My EA's test results - what's the catch and illusion? - page 4

New comment
 
Sta2066:
For 15 quid + traffic, there are plenty of them. I won't give you a link. There have been some problems lately. Gotta figure it out. But there's miles to go.


Yes, there are "millions" ahead))).... but as a Russian I want to get it for free... so I don't feel bad about it later)))) ... I've already found one for 400 rubles a month. I'm also looking for free ones abroad, I think I can find free ones there.
 
C-4:


Imagine you have an entry signal, let's say a buy signal. What does this signal mean? That the probability that the price will rise over a certain period of time will be higher than the probability that it will fall. If you set a stop, you are contradicting your own signal, instead of buying, when the probability of price increase is higher, you are selling - i.e. executing your stop. Therefore the stop itself has at best a zero mathematical expectation, and at worst a negative expectation (because your opposite signal is positive). And any test using abstruse trawls and stops shows this.


i can't imagine making money without stops.... i think it's the way to nowhere)
 
concord99:

I can't imagine making money without stops.... I think it's a road to nowhere.)
I'm not saying stops are evil. It's just that stops should improve the result of the TS - and that's hard to achieve. And any system must have rules for exiting profitable and losing positions. But those rules do not necessarily have to be stops and takeprofits, in the classical sense of the word.
 
TheXpert:
I've started to use time-based output almost everywhere instead of TA.
But do you still use stops?
 
DYN:
But do you still use feet?
Of course.
 
YOUNGA:

Maybe we should make a branch - output by TP, time, trawl (because I may have struggled in vain - a tricky trawl)

Time, trawl, trawl - this is bullshit. If a trusted TS gives signals for input, what prevents it from giving signals for output?
 
Figar0:

Time, trawl, tp - this is bullshit. The output should also be by signal. If the trusted TS gives signals to input, what prevents it from giving signals to output?
Just because a system has a signal to input and a signal to output, it is in fact two complete systems giving a signal to input.
 
C-4:
The fact that a system with an input signal and an output signal are in fact two complete systems giving a signal to the input.


Such TCs would be strongly interconnected... This is unlikely to work.

 
Figar0:

Time, trawl, tp - this is bullshit. If a trusted TS gives signals for entry, what prevents it from giving signals for exit?
Well, why? For example, if in a trend system the time of movements shows more stationarity than the size of the movement then it's more appropriate to exit by time rather than by TP.
 
Figar0:


Strongly such TCs would be mutually interlinked... This is unlikely to work.


An output signal that increases the MO of the system compared to a system with a random output gives an input signal in the opposite direction with a statistical advantage.
12345678
New comment
Reason: