History repeats itself - a lie? - page 6
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
Thank you, Alexey.
But the average number of repetitions (about 42) is catastrophically low for statistical analysis.
Maybe it makes sense to go down to H1 TF and go through the most volatile market hours,
for example, from 10.00 to 20.00 MSK? The amount of data in the sample will be much greater.
If we remove dependence on the day of the week, the data will be 5 times larger. I.e., an average of 600 repeats
of combinations - already something to work with.
Thank you, Alexey.
But the average number of repetitions (about 42) is catastrophically low for statistical analysis.
Maybe it makes sense to go down to H1 TF and go through the most volatile market hours,
for example, from 10.00 to 20.00 MSK? The amount of data in the sample will be much greater.
If we remove dependence on the day of the week, the data will be 5 times larger. I.e., an average of 600 replications
of combinations-- already something to work with.
If I were to perform such a research, I would only work with equi-bars.
Result not compiled, because 0 ... 0.4 ... 0.6 (0.56) is not interesting.
If I were to do this kind of research, I would only work with equi-bars.
Alexei (Mathemat) has just the right idea for the bars, you can develop it.
Aleksey (Mathemat) has a very suitable idea for bars, we can develop it.
IHMO there is no idea there (in the analysis of bars). And I haven't seen Alexey (Mathemat) claim it. The only thing that can be done is to increase the event frequency (not probability) using equi-squares but it's not a fact either, it's just a small theoretical assumption that may not come true.
Good luck with the research.
IHMO there is no idea there (in bar analysis). And I haven't seen Alexey (Mathemat) claim that. The only thing that can be done is to slightly increase the frequency of the event (but not the probability) by using equi-samples, but that's not a fact either, just a small theoretical assumption that may not prove to be true.
Good luck with the research.
Sergei, that's what I'm saying, just checking, and claiming is not my thing...I'm learning myself.
By the way, about probabilities - it's very hard to calculate, if possible at all, but here (the markets) the inertia component works and nothing else (in my opinion) is needed for successful trading.
I also searched for patterns by a (usual) candlestick code - on buy/sell I tried both mustached and bald and pregnant with hanged and married them in every way - but alas - the statistics is about 0.5.
Result not compiled, because 0 ... 0.4 ... 0.6 (0.56) is not interesting.
Can you elaborate on your approach to whether the event was successful or not?
Thank you, Alexey.
But the average number of repetitions (about 42) is catastrophically low for statistical analysis.
Maybe it makes sense to go down to H1 TF and go through the most volatile market hours,
for example, from 10.00 to 20.00 MSK? The amount of data in the sample will be much greater.
If we remove dependence on the day of the week, the data will be 5 times larger. I.e., an average of 600 replications
of combinations, that's already a lot to work with.
Yes, that's right, Alexander. That's what I wanted to write about in the https://forum.mql4.com/ru/14420 thread, but I decided to wait for the results of StatBars. As you can see my suspicions are confirmed. Even taking into account that "Director Foreh" supposedly doesn't give any super-secret information, the frequency outliers can't be considered statistically significant; they are just frequency fluctuations of very rare events, which can't be considered as evidence of dependencies... er... extremely reckless. If it were a ratio of, say, 510 to 370 rather than 51 to 37, significance would appear.
P.S. I'm going to be chastised again by granit77 for using swear words...
Yes, that's right, Alexander. That's what I wanted to write about in the https://forum.mql4.com/ru/14420 thread, but I decided to wait for the results of the StatBars script. As you can see, my suspicions are confirmed. Even taking into account that "Director Foreh" supposedly doesn't give any super-secret information, the frequency outliers can't be considered statistically significant; they are just frequency fluctuations of very rare events, which can't be considered as evidence of dependencies... er... extremely reckless. If it were a ratio of, say, 510 to 370 rather than 51 to 37, significance would appear.
P.S. Here we go again, granit77 will reproach me for using swear words...
I'm not a script, I'm quite a live person... ))
I must have said somewhere that the combinations can be used as a filter, but not as a strategy...
Let's see if Saltunov's system works together with several interrelated pairs...
2 Korey I would like to know the answer to my question (see above)...