Machine learning in trading: theory, models, practice and algo-trading - page 2711
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
When the "event" comes (becomes obvious or reflects in the past calendar) it is too late to rush and trade something somewhere.
So it can be just a signal to close, it is the same sales, if there were purchases.
In your picture "the price has reached ATR D1" is the consequence, the final, serious people have made all purchases/sales, then you can trade only noise.
Often yes, but it can be influenced by some other external event, let's say we are trading oil, for example, news - tanker is stuck and oil will not arrive on time. Here ATR is not so important....
Or purely technically - the market comes out of flat - every day the movement grows and ATR breaks through 100%, and with accompanying news it goes without pullbacks, i.e. it is possible to open further.
These nuances will be searched for in the second stage, when additional conditions to stumps will be built, where the usual behaviour when ATR reaches 100% will be revealed.
That is, the scheme kind of corresponds to reading in reverse time, from right to left : "price is going somewhere" -> "bought/sold" -> "damn, zigzag turn" -> "news, emotions" :-) .... but hypothetically it can be used by feeding quotes into it also from right to left, then the revealed consequences are close to the sought predictors and some of them can even be used.
Often yes, but it can be influenced by some other external event, let's say we trade oil, for example, news - tanker is stuck and oil will not arrive in time. Here ATR is not so important....
the news about "mega tanker stuck" is a force majeure (only if it is not periodically stuck in the same place and nothing foreshadowed). Immediately out of the market :-) Are you sure that the price will go up? and petrochem will get oil from the reserve (force majeure) and the price will fall.
The news about "mega tanker stuck" is a force majeure (unless it periodically gets stuck in the same place and nothing was foreshadowed). Immediately out of the market :-) Are you sure that the price will go up? and petrochem will get oil from the reserve (force majeure) and the price will fall.
This is just an example of the influence of information cancelling a condition that is more often fulfilled, ie just a split tree, that's what I was trying to say here.
This is just an example of the influence of information cancelling a condition that is more often met, i.e. just a split tree, that's what I was trying to say here.
You've never (maybe once) had a tanker get stuck before. You can't even guess whether or not there will be a split or what the reaction might be. The whole president of the USA, not the last informed man, thought that the quid would be at 300, but the gas came out at 3500 :-)
you've never (maybe once) had a tanker get stuck before. You can't even guess whether there will be a split or not or what the reaction might be. The whole president of the USA, not the last informed man, thought that there would be a quid at 300, but gas came out at 3500 :-)
So ideally there should be a split about an important event - about which everyone is trumpeting, the outcome is unknown in the sheet - for the model will be an opportunity to miss the signal to enter the market. We are not talking about the model in the example, but about the influence of some external events that cancels often repeated rules.
Ideally, there should be a split about an important event - about which everyone is trumpeting, the outcome is unknown in the sheet - for the model will be an opportunity to miss the signal to enter the market. We are not talking about the model in the example, but about the influence of some external events, which cancels often repeated rules.
Some external events can be divided into 1) about which someone knew something and these precursors are in the price 2) about which information is lacking or cannot exist at all. Any decision based on #2 is like "flipping a coin" - the system may decide "best case scenario A" but it is random. ; #1 can also be divided into 1.1 events that were not a surprise to you, and 1.2 all others. .... 1.2 is basically the same as #2, except that they can be minimised by self-development.
i.e. you can only include in the model what is known in advance, but perhaps not quite accurate and needs to be refined. It is impossible to answer the question "what to do and do something if the tanker suddenly gets stuck" on the basis of analyses of the past. Because it has not happened before.
some external events can be divided into 1) about which someone knew something and these precursors are in the price 2) about which information is lacking or cannot exist at all. Any decision based on #2 is like "flipping a coin" - the system may decide "best case scenario A" but it is random. ; #1 can also be divided into 1.1 events that were not a surprise to you, and 1.2 all others. .... 1.2 is basically similar to #2, except that they can be minimised by self-development.
i.e. you can only include in the model what is known in advance, but perhaps not quite accurate and needs to be refined. It is impossible to answer the question "what to do and do something if the tanker suddenly gets stuck" on the basis of analyses of the past. Because it has not happened before.
Everything can be put into an ideal system, but how much information we can get and process correctly is a question of realisation and abilities.
Everything can be put into an ideal system, but how much information we can get and process correctly is a matter of realisation and ability.
This is akin to the idea of piling everything into one pot and then pulling out a piece of gold. The main thing is to stir it the right number of times with a ladle of the right configuration :-)
it is akin to the idea of piling everything into one pot and then pulling out a piece of gold. The main thing is to stir it the right number of times with a ladle of the right configuration :-)
That's right, that's why I decided to group the data by making intermediate predictors by Events and then train on them.