Is it necessary to analyse the bars/curves on the timeframe? - page 5

 
Karputov Vladimir:

What I meant to say was that prices come in a price stream. This flow could be something like this (it's essentially a clipping from a tick chart):


But when this stream is rolled up into timeframes - you get completely different views/figures (it all depends on your imagination) on different timeframes. But collapsing data into timeframes kills something in the data. Something very important.

As rightly pointed out here, yes it does, it kills the high-frequency component of useful information. However, there is still low-frequency information.
 
Also in statistics, max and min values are usually discarded, and in a bar this is half of the price information.
 
ratnasambhava:
The market is biased - the second tenet of thechanalysis.
So, abrupt but not quite.

What do you mean by "abruptly, but not quite"? Can you give examples on a price chart?

 
Andrey Dik:

Delay is not the worst thing from filtering (smoothing), the worst thing is the price traded, not the result of smoothing.

In short, yes, a dead end. It's more effective to deal with reversals in other ways than trying to transform price and inventing "advanced candlesticks".

Could you be more specific? Because I, too, may speak of other ways, more advanced methods, which should be improved and deepened))).

We trade the price and determine the direction after the filter, what's unclear?

 
Alexey Volchanskiy:

1. Could you be more specific? Because I can also talk about other ways, more advanced methods, which should be improved and deepened ))))

We trade the price, and the direction is determined after the filter, what is unclear?

To be more specific:

1. The classical thechanalysis uses two inherent price properties directly related to its discreteness - the trend movement (when consequently, each candle is higher/lower than the previous one), and the rebound from the level / reversal.

Both these properties can and should be successfully exploited. For example, levels, if the price is currently in the area of some accumulation of historical prices, it is reasonable to reduce the position size many times (down to 0), if there is a signal to enter. This is the best-known and most widely used method of dealing with "a sharp change of direction". There are others - the channels of slope, for example. There are also more advanced methods (neural networks and other methods of machine learning).

2. That's the problem, what's not to understand?

 
Andrey Dik:

More specifically:

1. Classical thechanalysis uses two intrinsic properties of price directly related to its discreteness - the trend movement (when consistently each candle is higher/lower than the previous one), and the rebound from the level / reversal.

Both these properties can and should be successfully exploited. For example, levels, if the price is currently in the area of some accumulation of historical prices, it is reasonable to reduce the position size many times (down to 0), if there is a signal to enter. This is the best-known and most widely used method of dealing with "a sharp change of direction". There are others - the channels of slope, for example. There are also more advanced methods (neural networks and other methods of machine learning).

2. That's the problem, what's not clear?

The problem is that we're back to the ancient candles again)) OK, the conversation is pointless, we are speaking different languages.
 


You have to simplify the logical chain :) The herds go to the mountain, and only after that do you help them, not the other way around. You don't drive them there, you help them since they went there.
 
Alexey Volchanskiy:
The problem is that we're back to the ancient candles again)) OK, there's no point in talking, we're speaking different languages.

It's not about the candles, it's about the price discretion. You and the top-trader do not take that point into account. Renko-like constructions are not without the disadvantages of classic OHLC candlesticks, and even contain less information.

The only thing I would suggest as part of the topic is to try to come up with "advanced" candles, you need to add information to OHLC. And just OHLC are already really, to put it mildly, outdated.

About "add information" - I don't have any ideas yet, there's a lot to think about specifically on this subject.

 
Andrey Dik and Alexey Volchansky, forgive me for my immodest question. With such hyperactivity on the forum, do you have time to write articles you've subscribed to? Will they be published this year?
 
Yuri Evseenkov:
Andrey Dik and Alexey Volchansky, forgive me for my immodest question. With such hyperactivity on the forum, do you have time to write articles you've subscribed to? Will they be published this year?
I speak for myself - they will. I can even tell you a little about "sweetness" - I have developed and tested a new statistical indicator, which replaces the usual PF (True PF - TPF). There will be other "goodies" too.
Reason: