From theory to practice - page 16
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
Andreturn increments are indeed non-stationary .
But that doesn't mean that you can't work with them because of that. The only question is, to work for what result?
What is the goal here? To write a market formula? Or to remove Cash? And these goals are very different.
Oleg, I also have a question for you - what is the point of working with each tick? Why do traders here fight so hard for it?
I personally think it makes no sense.
And "Why traders are fighting so hard for this? -- is a way of perceiving reality.
What is the point of working exactly with each tick? Why do traders here fight so hard for it?
Because a tick is a quantum of the trading world, the smallest indivisible particle from which any timeframe is sliced together. And any type of chart: Renko, Kagi, three-line breakout, crosses zero, bar. In other words, we can already see that the chart is an indicator of the price, but the price itself is a tick.
But in your reasoning they can be neglected, if you build a model based on prices measured at equal periods (for example close prices), and volatility over the measurement period (price high and low), and volatility is separate, separately from the high and low. Or you can calculate the midline of volatility and the drift of closing from it (the midline of volatility).
If you build such a model, you may not look at ticks.
And by the way, what is the physical pricing model? How do you imagine it? Maybe if you philosophize in this direction, then an adequate mathematical model of the process will appear.
We all understand that a stream of orders flies into the market and matches the liquidity, so what is the nature of this flow? The fact that this flow has a memory is incontestable, but does it have inertia? Again, what kind of memory? The trader analyses not the quotes but the quotes, so he reacts to the approach to certain levels, etc.
Because a tick is a quantum of the trading world, the smallest indivisible particle from which any timeframe is sliced together. And any type of chart: Renko, Kagi, three-line breakout, crosses zero, bar. In other words, we already see that the chart is an indicator of the price, but the price itself is a tick.
But in your reasoning they can be neglected, if you build a model based on prices measured at equal periods (for example close prices), and volatility over the measurement period (price high and low), and volatility is separate, separately from the high and low. Or you can calculate the midline of volatility and the drift of closing from it (the midline of volatility).
If you build such a model, you may not look at ticks.
And by the way, what is the physical pricing model? How do you imagine it? Maybe if you philosophize in this direction, then an adequate mathematical model of the process will appear.
We all understand that a stream of orders flies into the market and matches the liquidity, so what is the nature of this flow? The fact that this flow has a memory is incontestable, but does it have inertia? Again, what kind of memory? After all, the trader analyzes not the quotes but the quotes, so he reacts to the approach to certain levels, etc.
Each tick should be considered only because it is ticks that are sent to the MT5 input. And no timeframe or bar does not play any role here.
Imagine that this is a football game. Can you play football based on some "bars"?
Imagine a football player stands and waits for a certain time and starts to act when the pause time is over, i.e. when a bar is formed. Isn't it funny?
That's why you have to consider every tick and act according to the situation, because every next tick will be able to change the situation.
And if, with the development of technology, tics start to come in at frequencies of up to 1 microsecond, then they should ALL be accounted for too????
With the development of technology you can easily account for them and process them!
In the 16th century, the Japanese didn't have technology, and worked with only 4 numbers in a day, calling them candles.
And now we're getting to ticks...
And if, with the development of technology, tics start to come in at a frequency of up to 1 microsecond, then they should ALL be accounted for too????
I may surprise someone, but this is already happening, without waiting for the development of the technique. The real trades are much more frequent, but brokerage companies build filters and send already filtered, brushed datafeeds to the tick-feed.
For example if you do not know what to do with a robot, it may be a good idea to start a new one.
So yes, you have to build robust model, not depending on tick statistics. By the way, the change of the filter can be traced on the M1 history through the change of tick volume statistics.
The data will be used as a basis for the analysis, but we will not know the secrets of the algorithms. He just mentioned that you can't trade on uncoordinated datafeed, you'll get stuck with requotes.And if, with the development of technology, tics start coming in at a frequency of up to 1 microsecond, then they should ALL be accounted for too????
The human brain is capable of processing information faster than the most powerful computer. It has something that no computer has, and that is intuition and ingenuity.
So it's the other way round - chasing every single tick is completely unnecessary and it's OK to accept ticks at a certain interval if we're not arbitrating. Am I getting this right?
Not exactly, statistics must be gathered (it's important data), but the flow itself changes when the filter changes, and it is found out usually post factum.
The problem of physicists is that all their equations are tied to delta t, so the frequency generator must have clear frequency, but with ticks it floats. The only price we can rely on in terms of time is the bar closing price. In this case we know for sure that the bar will close at that time.
And even in close we have gaps in time on weekends.
Exactly, Nicholas! But a bar is ticks collected over a minute, and if we imagine some sort of seconds bar, isn't it right to think that the last tick arriving within a second is what we need???
The average tick rate in many DCs is 3 - 3.5 sec, by setting the parameter to 1Hz you are going below the flow sampling rate.
Even with a 3Hz sample rate, we're getting unstable flow. You can only ignore the instability from 30sec onwards.