Demonstrating the cluster approach to market... - page 29

 
granit77:
Nothing to be seen from the front, nothing to be seen from the back (c)
What period of optimisation? Is there a forward on the screen? Martin, or constant lot? Entries by crossover only, or is there a fill by third-party signals? Closes by counter signal?

No optimization, lot is fixed, entries only by crossing, closes by counter signal or break-even.
 
trol222:

This may require considering the price movement from one TF to another (roughly speaking). The amplitude of this movement will not reach the level of the phase of a higher TF and the sell signal will not be present

That's right. That's how it's done. Of course, you can do it on one TF, but why strain the processor?

Vitya:
You can try to use crossover of indices (x,y), for entry signal to catch long trends. Exit by the opposite signal.
I used to do so. The signal is good on a trend. During the flat the profit is zero (the signal lags). I have switched to synchronous reversal. It always works.

 
Zhunko:

That's right. That's how it's done. You could, of course, do it on one TF, but why strain the processor?

I used to do so. The signal is good during a trend. During flat periods zero profit (the signal lags). I have switched to synchronous reversal. It always works.


i.e. signal received on H4 - waiting for confirmation on D1?
 
Vitya:

i.e. signal received on H4 - waiting for confirmation on D1?

Reverse: D1 -> H4.

 
Zhunko:

Conversely: D1 -> H4.


If we consider the previous figure in this context:

If a buy signal is received on D1, only signals that do not contradict the main trend are accepted on H4.

 
Vitya:


If you consider the previous figure in this context:

If there is a buy signal on D1, only signals that do not contradict the main trend are accepted on H4.

With the previous composition of indices we will not see these signals, before writing buy, we need to cut the bottom couple off and replace them.
 
marketeer:
With the old line-up of turkeys we won't see these signals, before writing the buy, we need to cut the bottom couple off and replace them.


It is also possible to consider only those sections, where x rises and y falls at the same time (H4). The same situation is on (D1)

 
Vitya:


It is also possible to do the same with these, but only count the sections where x increases and y decreases at the same time (H4). The situation is the same on (D1)

Not exactly. The idea was to analyze the yield at the overbought (X) and oversold (Y) extremes and if we look at the internal "declines" with further widening, there will be a lot of false signals. Anyway, I have to start the lab work here on Monday in practice. ;-).
 
Vitya:


If you consider the previous figure in this context:

If there is a buy signal on D1, on H4 only signals that do not contradict the main trend are accepted.

You can, for example, do the following: W1 -> D1 -> H4 -> H1, etc. ....

The more confirmations, the stronger the signal, but the number of signals decreases. You can thus make a TS with a profit factor approaching infinity.

 
Zhunko:

You could, for example, do the following: W1 -> D1 -> H4 -> H1, etc.....

The more confirmations, the stronger the signal, but the number of signals decreases. You can thus make a TS with a profit factor approaching infinity.

Yeah, and the time between trades will also tend to infinity ;-).