Machine learning in trading: theory, models, practice and algo-trading - page 2635
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
What do you mean by regular such samples?
every day, every week - the exact same thing happens. Carefully take a day by day view, M5-M15 resolution is sufficient.
Also the reaction to the news. You open your trading log and from the marked serious events, an hour before and 3-4 hours after, in minutes - neural network feeds and MO
Same with regular events, for example fixing twice a day for gold and once for euro. You know, there is an event with consequences. You must understand the consequences and the events are regular, you do not need to predict them.
every day, every week - the exact same thing happens.
Is that a statement?
that's how it really is
that's how it really is
I've tried the same as you described, or nearly the same, but I have not found any regularities that would indicate that "every day is the same.
wrote a long reply, pressed "add" and the site crashed: nginx's favourite Bad Gateway happened
It's a hint from above - not everything has to be told :-) look it up, it's there. It's not so much a grail, but the quotes are no longer perceived as noise and random.
These "practitioners" with their "practical eye") They "see" waves, or they see "scheduled price movements").
Fluctuations/volatility spikes are visible and even partially predictable (a long known fact). Directional moves are much less predictable. That said, volatility fluctuations greatly spoil the ability to find the latter.
These "practitioners" with their "practical eye") They "see" waves, or they see "scheduled price movements").
Fluctuations/volatility spikes are visible and even partially predictable (a long known fact). Directional moves are much less predictable. At the same time volatility fluctuations greatly impair the ability to find the latter.
Oh, those "theorists". :-)
Practical methodology for working with real time data:
- We take into account winter-summer time transitions, i.e. we calculate statistics for winter time separately and for summer time separately. Specifics of time switch will make 2-3 weeks disappear from analysis. Then we sum up
- we choose the reference point for the periods. Because 0:00 is good for nothing - it introduces a lot of noise. Define criterion, look. I've got an optimum of 4 hours. Let's call it Nemo's point.
- and at the same time decide what to do with Monday and Friday. The candlesticks will be a little more "dense". You can't tell much by eye, but it's statistically obvious.
- Let's classify into "anxious", "normal", "liquid/weak market" on the histogram H-L. 10+80+10=100% or other percentages of your choice
- We put "liquid market" on the calendar and compare it with various holidays and weekends in the financial world. They overlap, almost perfectly. Ok - we know in advance when there will be a liquid/weak market and we have our own strategies and analysis for it.
- the same trick does not work with "worrying" days.
- We pull out everything we can about "normal days" - statistics, rates, deviations from Nemo. We get some areas where a reversal is most likely and tasty.
- For example, we teach the neuronics to distinguish the onset of an "alarming day". There is no need to trade - just let it whistle "Master, something is wrong" on time.
No, not a universal grail, but you can trade with this one and enter the market. Then we remember that there are other currencies and they are all interrelated and so on....
These "theorists". :-)
Very strange description and very strange terminology
liquid market is a normal, common term.
"Alarming days" is, of course, a novelty. That is why they cannot yet be called trendy or volatile.
the liquid market is a normal, conventional term.
"Alarming days" is, of course, a novelty. That is why they cannot yet be called trending or volatile.