From theory to practice. Part 2 - page 50
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
The histogram of the time intervals must satisfy an Erlang distribution of a certain order. For each pair, a different distribution.
Only the Absolute time of the system according to Poincaré remains unchanged - day, week, month, year, etc.
I was thinning out eurusd on some minutes (with the help of some algorithm I cannot remember now) a histogram similar to this one was obtained:
My peak was about 3 minutes.
I was thinning the eurusd on the minutes (using some fictitious algorithm I can't remember how) and I got a histogram similar to this one:
My peak was about 3 minutes.
Something similar...
But, I don't work on minutes. I work on seconds. I get the same distribution at 3s and I use it. No grail, of course, but something like a paltry profit is there.
Rena, I can't help but reply to you.
Hysteresis in the market is a phenomenon that is being massively investigated. Some things I've even been banned from posting on the forums. Like isocline :)))) There's a whole cult out there and I don't want to get into it. They may be having successes - I'm happy for them.
From my point of view, the fact that hysteresis is on the market says it's non-Markovian. That's good enough for me.
Rena, I can't help but reply to you.
Hysteresis in the market is a phenomenon that is being massively investigated. Some things I've even been banned from posting on the forums. Like isocline :)))) There's a whole cult out there and I don't want to get into it. They may be having successes - I'm happy for them.
From my point of view, the fact that hysteresis is on the market says it's non-Markovian. That's good enough for me.
Also.
when I was looking for a way to calculate hysteresis, I wrote down what I found on the web into this source
here is a beginning on AB=BC and text on how to make hysteresis via sines and cosines
https://www.mql5.com/ru/forum/366487/page2#comment_21674350pf=1.5, average trade 15p
It's almost the same on the minutes. No thinning, of course, is needed here.
By optimizing the window and the multiplier you can find a better variant, but it is surely a fitting.
The same situation is in USDCAD and the situation at other pairs is worse.
Visual inspection of trades suggests that the system does not "see" any returns and moreover trends.
By the way, here is a test of Shurik's system, on EURUSD Open H1, for three years 2018-2021, with a spread of 1p:
pf=1.5, average trade 15p
It is almost the same on the minutes. No thinning, of course, is needed here.
By optimizing the window and the multiplier you can find a better variant, but it is surely a fitting.
The same situation is in USDCAD but the situation at other pairs is worse.
Visual inspection of trades suggests that the system does not "see" any returns and moreover trends.
36 deals, an average of one a month
not much
By the way, here is a test of Shurik's system, on EURUSD Open H1, for three years 2018-2021, with a 1p spread:
pf=1.5, average trade 15p
It is almost the same on the minutes. No thinning, of course, is needed here.
By optimizing the window and the multiplier you can find a better variant, but it is surely a fitting.
The same situation is in USDCAD and the situation at other pairs is worse.
Visual inspection of trades suggests that the system does not "see" any returns and moreover trends.
The idea was to select a set of parameters on a smaller timeframe for trading on continuation of movements (persistence), and on a larger timeframe - for trading on reversals (antipersistence), like in Shurik's case. This partly corresponds to the behaviour of currencies on the average - continuation of small movements and reversal of large ones. Then we could try to build a portfolio of several such systems with different parameters.
Obviously, due to non-stationarity, which is clearly visible in your picture, everything would have turned out the same way as usual) There is too much to do to try to bring Shurik's idea to life. Let him, let him do it himself)
The idea was to select a set of parameters on a smaller timeframe for trading on continuation of movements (persistence), and on a larger timeframe - for trading on reversals (antipersistence), like in Shurik's case. This partly corresponds to the behaviour of currencies on the average - continuation of small movements and reversal of large ones. Then we could try to build a portfolio of several such systems with different parameters.
Obviously, due to non-stationarity, which is clearly visible in your picture, everything would have turned out the same way as usual) There is too much to do to try to bring Shurik's idea to life. Well, leave him to his own devices)
You'd think any other idea/system would require less fiddling...
The idea was to select a set of parameters on a smaller timeframe for trading on continuation of movements (persistence), and on a larger timeframe - for trading on reversals (antipersistence), like in Shurik's case. This partly corresponds to the behaviour of currencies on the average - continuation of small movements and reversal of large ones. Then we could try to build a portfolio of several such systems with different parameters.
Obviously, due to non-stationarity, which is clearly visible in your picture, everything would have turned out the same way as usual) There is too much to do to try to bring Shurik's idea to life. Well, let him do it himself).
I also wanted to do the analysis of movements on small TFs, e.g. 1 minute, with the purpose of predicting bar characteristics on a large TF, e.g. a day. I.e. try to predict Hg, Lw, and direction Cls-Opn, say, of a day candle by previous movement on a minute TF. And then somehow refine the prediction based on analysis of the daily chart. It is clear that one cannot gain a great advantage, but something acceptable may be obtained using a portfolio on several symbols. But I have not got down to it yet.
continuation of small movements and reversal of large ones.
The problem is that small moves are more often reversed and large ones in a trend take away most of the profits of small reversals.