Looking for patterns - page 126

New comment
 
Aleksei Stepanenko:

All right, but still, someone lives there.


A graph of random increments:



Real graph:



Did you form the random increments with a fair coin? If so, you got a Gaussian. Take a thick-tailed distribution and all the pulses will be no less than in real quotes, although these are also random increments. Or in the simplest case you can randomly change the expectation of a coin on some interval.

Market quotes are random increments with changing statistical characteristics. The only good thing about all this is that the statistical characteristics may not change very quickly.

 
sibirqk:

Did you form the random increments with a fair coin?

Yes, pseudo-honest. A macro in Excel:

Cells(1, 1) = 1
Cells(1, 2) = 100
For i = 2 To 5000
    Cells(i, 1) = i
    a = 0.1 * (Round(Rnd(1), 2) - 0.5)
    Cells(i, 2) = Cells(i - 1, 2) + a
    Cells(i, 3) = a
Next i

But without taking volatility into account. Therefore there were no visible strong movements. However, if you add daily volatility, I think everything will be fine, both movement and flatting will appear! Just like Genghis in the picture.

 

The regularity is that if we trade in the trend, the baika (or most of them) may end up at the hai, and the sellka (or most of them) may end up at the low.

//It is not very difficult for the market (quotes) to do this.

Together both trades will create a negative lock that perfectly fits the market, because negative profit will concentrate inside it, i.e. the buy and sell trades will fall in minuses simultaneously.

If counter-trend trading is considered, it is vice versa and the market does not benefit from it.

The quotes in this case have only one choice - the black swan or the long "no-back", while the trader has to reduce the risk.

 
Renat Akhtyamov:

then the baika (or most of them) may end up on the hai and the sellka (or most of them) in the loi.


Yes, but there doesn't seem to be a powerful dolt, it's the essence of the market, a self-directed system. Otherwise it would be a printing press for all, which is impossible.

 
Aleksei Stepanenko:

Yes, pseudo-honest. A macro in Excel:

Cells(1, 1) = 1
Cells(1, 2) = 100
For i = 2 To 5000
    Cells(i, 1) = i
    a = 0.1 * (Round(Rnd(1), 2) - 
    0.5)
    Cells(i, 2) = Cells(i - 1, 2) + a
    Cells(i, 3) = a
Next i

But without taking volatility into account. Therefore, no visible strong movements occurred. However, if we add daily volatility, I think everything will be fine, both movement and flatulence will appear! Like Genghis in the picture.


I'm not very good at Excel macros, I prefer Matlab, but apparently the selected figure should be randomly changed to - 0.4, 0.6, 0.5, say, once an hour. Then the interval of quotes with given expected payoff should also be changed randomly - for example, one hour plus minus half an hour. And finally add the modulation of daily volatility. Then we will obtain something similar to market quotes at a very first approximation.

 
Spread, swap, and commitments reduce the probability of a random win. Spread, swap, and commitments are the dolt.
 
sibirqk:
Then you get something in a very first approximation similar to market quotes.

Yes, you are right. You could do hourly volatility statistics over a 24-hour period, so you get a spread over each hour. Then randomly take a number from the range of that swing, and... - real EURUSD chart in Excel or Matlab

 

Not to patterns, just maths.

Rather, it's already known.

Candle[Close - Open] = ((Ask_Point_Upper - Bid_Point_Lower) + (Bid_Point_Upper - Ask_Point_Lower)) / 2

 
Evgeniy Chumakov:

Not to patterns, just maths.

Rather, it's already known.

Candle[Close - Open] = ((Ask_Point_Upper - Bid_Point_Lower) + (Bid_Point_Upper - Ask_Point_Lower)) / 2

cool formulae

and why count the average, wouldn't it be more useful to compare, i.e. subtract?
 
Renat Akhtyamov:


and why count the average, wouldn't it be more useful to compare, i.e. subtract?


Try it.

1...119120121122123124125126127128129130131132133...306
New comment
Reason: