Scold :) Interested to hear your opinion regarding... - page 15

 
YuraZ писал (а) >>

are you coming out to 2008 with it?

Don't know yet ...

Don't even have the rules yet :(

 
Mischek писал (а) >>

it's a pity...

the word "exercise" you've either forgotten...

I'm too lazy to start again.

Maybe I'm imagining things, but if I ask you about the weather tomorrow you'll reduce the conversation to a "trivial over-sitting".

No, I haven't forgotten about the exercise. I have an intuition. So what?

About the weather, so what if I can guess and think I'm guessing? Until statistics show that I am guessing correctly with a certain probability, there is no use in conceiving.

 
Lovecraft писал (а) >>

Increases after a loss, when the probability of taking a profit is almost 100%. I am trying to get the Expert Advisor to trade almost every day.

Question to connoisseurs: can I trust quotes from a 3-4 year old quote archive?

You would be doing everyone a great favor if you tell us why all of a sudden: "after a loss when the probability of withdrawal is nearly 100%" Where did you get this knowledge? How did you get it?

 
ForexHelp писал (а) >>

You've got it all wrong. What does this have to do with over sitting? - You wrote that you were ready to open on a bad signal. Maybe I don't understand why your signal is bad. If it is very likely to produce a small profit, what does it have to do with catching trends of 1500 pips? I understood that the signal is bad in the sense that it has a low probability of making a profit, so we open at random to wait out a small profit or a big trend.

If the system can catch trades of 1500 pips, and on average more than 100, but since you never know what will happen tomorrow, you can close at moments when "logic" says to close, and find the moment to enter again on the trend. Or if the signal is long, you may not catch 100-300 pips due to coming back. That's what I mean. - I understand that signals predict on how many points the trend will "shoot" or something like that, right?

Just to cut the profit and not to let it grow - it is trivial, you know, it is not even worth talking about it - it is clear to everybody. - It's not trivial. It is statistically justified, i.e. practically scientific. A formalized system with a calculated statistical advantage "just chops profits" with fixed tees and is in full statistically justified confidence that this approach maximizes the overall pie, i.e. to make the overall profit grow - chop profit in each trade.

I am also familiar with the approach you are probably referring to - a system that predicts the beginnings and ends of trends. Nothing fixed, like following the market, taking the maximum from the trend. Any forum is full of beautiful examples and descriptions of how it should be done. Examples are given for segments with a clear trend, the stories do not provide statistics, but only magic formulas of following the market and the magic MM. As a rule, the tale ends when we are asked to give the test results for a longer period of time, where there was everything, not only a convenient trend.

And about the losses - the logic is that when the signal appears this position is closed and opened in the other direction, as a result the losses are minimal. - This statement becomes true if the signal has a statistical advantage in its correctness, otherwise, this claim of catching peaks (swings) is a trivial manipulation, of which there are countless and can be created from morning till night. Do you have statistical evidence that your signals are pointing in the right direction?

Here's an intermediate for 2008 (there's a big drawdown right now, but it's such a working version. Basically, I already did it at 4.5%, and PF 2.0, it works with one lot as main, and one for scaling in the trend sometimes):

And also here averages multiply by two, because in fact it is torn off simultaneously 0.1+0.9 (for convenience), therefore averages are such. It turns out that $2000 average per lot of profit, and 1000 loss. Is it an overbetweening? :) - Average profit and average loss says nothing about how much maximum drawdown your working version is willing to outlive. Tell me what stoploss your system has and it will become clear.

 

Vita, I repeat my question.

Yurixx писал (а) >>

2 Vita

I like your approach, I have a similar point of view.

It would be interesting to add something else to that about how to determine whether a TC gives a stat advantage and, if so, how to calculate it.

 
And once again I second the question
 
alexx_v писал (а) >>
And once again I second the question.

>> Yes, of course. In general, the answer is prosaic - use statanalysis. To do so, you have to learn the math and use it as intended. It is very important to be consistent and logical, otherwise all sorts of tricks and errors can add up to anything. In each particular TS, you can make a wrong logical assumption, or process data that is not relevant to the subject of the study, or you can do anything wrong.

But I sense that you are asking about something else, something specific?

 

Let's give a simple example. Weekly chart of the pound. At the opening we buy and wait for 5 pips in profit. How would you determine if there is a statistical advantage to this idea?

I get 1537 positive outcomes on 1591 bars, i.e. a profit of 7685 pips. Spread of 3 pips was taken for the whole period, which should give a bias towards a prettier picture.

Does anyone know how to calculate the loss?

 

В общем виде ответ прозаичен - воспользоваться статанализом.

Somebody, I do not know who or when or how, having learned the basics, gave out statistics that say that 95% of traders are losing and only 5% are earning. We will assume that this statistic is true, with some small, I would say floating, margin of error.

What is the statistical advantage of traders and their TS?

Roughly speaking, on a "smaller timeframe", say on the "hour", the trader's TS supposedly has statistical advantage, and on a larger "daily timeframe", 95% to 5%, the statistical advantage is not only absent, it is absent...

Is it a statistical advantage in the absence of a statistical advantage or not?

 
Prival писал (а) >>

I don't agree with this categorical statement. There are systems (at least theoretically) with 50/50 profit/loss trades. But they can be super Grails and precisely due to MM. These are sequences of PPPUUPPPUUPPU ...

P - profit, U - loss. I think it's clear how to use MM, but I don't think it's impossible to find such an ideal system but we should check this system to see if this regularity is present in the sequence of trades.

Example. A trend following system. There is a lot of small losing trades in the "flat" and hope for one that will catch the trend and block the losses. This can be modified as soon as a loss is incurred, the lot is minimized and as soon as a profitable trade appears, we begin to increase the lot. As if that ONE large profit is divided into many small but increasing lots + we attach the forecast (statistics of TC) by the length of a trend being captured.

How do you like it?

S.I. In TS should be all beautiful, soul, body and control.

No, it's not.

The mathematical theorem states that you cannot build a winning strategy when the outcome of a trade is 50/50. You, on the other hand, are simply asserting the opposite. Like you can make an MM which will produce trades with 50/50 outcome, but with profits and losses coming in a regular order. And, of course, it is not difficult to exploit this pattern. Only a person who has a very hard time with one-way logical conclusions can believe it. Of course, you can believe what you want, but the mathematical theorem does not care whether your system is theoretical or not, it does not care how many times and how you'll play with the results of transactions, because no sophistical tricks will not make it change its conclusion - you cannot build a winning strategy on a 50/50 split.

You gave an example of a common fallacy that turns a blind eye to the nature of the pattern in the PPPUUPPPUU sequence... Where do the legs of this pattern, on which it is so easy to build a winning strategy, grow from? From a 50/50 imbalance, but not from the magical properties of MM. First there has to be a statistical advantage in the initial series, only then will a pattern emerge in the sequence of trade results, otherwise we contradict the matteorem.