Twenty-four, and not one iota more! - page 12

 
Cod:


If it's a bank account, I'll take it more or less seriously. But if it's a jpg picture...

Well if you manage to pair a fixed stop with high volatility or for example a thin market, I'll consider that I've spent so much time picking guts out of charts ))))
 
artikul:

Well if you manage to pair a fixed stop with high volatility or for example a thin market, I'll consider that I've spent so much time digging guts out of charts ))))
Bravo. Put it in the anthology...
 
Cod:


This approach is life-shattering... By my calculations, intraday stop, technically reasonable - 65 pips. So how do you build profitable trading with this system?


Try to connect different variants of stop, followed by position support, for different TS with different TFs - different options, everything is strictly individual, check, test it ... Try adaptive c/l - by 2 APR - with different periods - the first - up to 10, the second - from 10 to 30; + multiplier (depending on the volatility of the market) choose the one that is larger and set. All optimized on the history, connect the best variant and go ahead - reconsider your attitude to optimization, including the work including stop on the history... see here - https://www.mql5.com/ru/forum/107064

In the attached file is a library of trailing stops - try it out, plug in the "best" one and go.

Files:
 
Crocodiles are useful animals.
Roman, the person in charge doesn't recognise history tests as a matter of principle.
 

So... Just thinking out loud. No desire to make a new branch.

Tester and optimizer.

I'll leave the implementation on the authors' conscience, but I wanted to say something else.

As in Metastock, as in Omega, as here - all the same. The tester is a way to find out where you screwed up. No more than that. And on mokla - especially because it's sharpened for trade.

Optimization is a weird thing. Teen alchemy.

If you have an idea, what does optimization have to do with it? If you don't - why should you ravage your CPU?

Dr. thing is that sometimes, running another test, you find out some unknown peculiarities. So as a research tool optimisation (process) will do.

That's all for now... )))

 

I do not believe in optimization on the history, because I know how two or three studs in a month can fundamentally change the whole picture and give an answer to the question "where to place a stop and where to take" with an error of nearly hundreds of pips. I.e. it is optimized for the most distorted prices that will probably never happen again. This is nonsense.


 
Cod:

I don't admit to optimising on history because I know how two or three stilettos in a month can fundamentally change the whole picture .....


How is it that a couple of extra stops can fundamentally change the whole picture?
 
Svinozavr:

So... Just thinking out loud. No desire to make a new branch.

Tester and optimizer.

I'll leave the implementation on the authors' conscience, but I wanted to say something else.

It's all the same in Metastock, Omega or here. The tester is a way to find out where you screwed up. No more than that. And on mokla - especially because of the sharpening of the trade.

Optimization is a weird thing. Teen alchemy.

If you have an idea, what does optimization have to do with it? If you don't - why rape the CPU?

Dr. thing is that sometimes, running another test, you find out some unknown peculiarities. So as a research tool optimisation (process) will do.

That's all for now... )))

Another way to build a non-verbalizable algorithm (the one that a trader/programmer can't formulate, but is sure it exists) or to check that such an algorithm can't be built using the selected methods using this source data ;)..... Alas, a "broader" implementation of the methodology is needed. And here we go into the field of AI.......

A standard tester is approximately like a primitive man in the chain of evolution (no offence to the developers - in principle, there is nothing more to be desired from the free software), but if you know how to use it, it allows you to cut out obviously dead-end ideas.

IMHO - sometimes the traders'/programmers' games with standard tester remind a phrase about a blonde driving a car or a monkey with a grenade ;).....

Good luck.

 
paukas:
How is it that a couple of extra stops can fundamentally change the whole picture?

I'm talking about studs. So, you trade quietly for a month, till the 29th day, everything is quiet and orderly, you get your stats, and suddenly on 30 and 31 sticks out a couple of these crazy spikes... And it turns out that your optimized TP, instead of 40 pips, should be 120 (or vice versa), in short, all the normal market conditions go sideways, and the whole point of optimization is lost, because you're not optimizing for the normal market trend, but for two crooked bastards. Drop the machines, run your hands a couple of times, at least for a month, and you will see the impact of completely random outliers on the overall result, on the TP-SL combination. So it's self-defeating.
 
Cod:

I'm telling you, it's a hairpin. So you trade quietly for a month, until the 29th day, everything is quiet and orderly, you get your statistics, and suddenly on 30 and 31 sticks out a couple of these crazy spikes ... And it turns out that your optimized TP, instead of 40 pips, should be 120 (or vice versa), in short, all the normal market conditions go sideways, and the whole point of optimization is lost, because you're not optimizing for the normal market trend, but for two crooked bastards. Drop the machines, run your hands a couple of times, at least for a month, and you will see the impact of completely random outliers on the overall result, on the TP-SL combination. So it's self-defeating.

That's what I'm asking. How can two cases affect 200? Reduce income by 2%? Is that a disaster?

Yes, and there is no such thing as a "normal market" and an "abnormal market".

Reason: