From theory to practice - page 718
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
Of course, I'm going to sound amateurish. But I will, because it's interesting to get to the bottom of it.
It turns out that the broker at a given moment in time, there is a certain array of prices, whose distribution satisfies the Poisson distribution. At some non-random(as it belongs to the k-th order Erlang distribution) interval of time, the broker outputs a certain value - a tick for the clients from this array.
That's exactly what we get Skellam's distribution for returnees.
Right?
The broker has:
- The price array is a buffer which is ruled by the tumbler. This is where the limits are (coming in/going out). In general terms, not far from the spread (the place where the "spreading" takes place, the filling is like a parabola with the volume ~ sqrt(x) ). In order for the price to pass X<STOPLEVEL, all of the limits along the way must be served. Limits beyond the Stop Level may go up or down in a dynamic way.
- there are streams of market orders to buy/sell at the current price. You can think of the orders as having the same volume, just coming in at different rates, independently of each other. Once in a while (and/or according to the accumulated volume, this is its discipline of service) the broker looks at what has accumulated, makes offsets, and uses the nearest volumes from the cup for what has not been set off. In its internal kitchen broker can use values less than a point, but if the price changes above the threshold TICKSIZE is generated TIC
This is of course simplified :-) there are also market makers who provide liquidity, i.e. they refill the market with limits around the spread and do not let bid-ask go too far apart. And for market flows it is better to consider that there are independent flows which have higher volumes.
enough to draw a pie chart with squares, output your distribution (result of objective measurements) and gradually make conclusions by refining model/flows/connections/distributions. One day the squares will cease to be totally black and you will be able to trade on the model. And you cannot trade just on any distribution
The broker has:
- The price array is a buffer, which is called a tumbler. That's where the limits lie (come/go). In general terms, not far from the spread (the place where there is "spreading", the filling is like a parabola with the volume ~ sqrt(x) ). In order for the price to pass X<STOPLEVEL, all the limits along the way must be served. Limits beyond the Stop Level may go up or down in a dynamic way.
- there are streams of market orders to buy/sell at the current price. You can think of the orders as having the same volume, just coming in at different rates, independently of each other. Once in a while (and/or according to the accumulated volume, this is its discipline of service) the broker looks at what has accumulated, makes offsets, and uses the nearest volumes from the cup for what has not been set off. In its internal kitchen broker can use values less than a point, but if the price changes above the threshold TICKSIZE is generated TIC
This is of course simplified :-) there are also market makers who provide liquidity, i.e. they refill the market with limits around the spread and do not let bid-ask go too far apart. It is better to think that there are independent streams with larger volumes.
enough to draw a pie chart with squares, output your distribution (result of objective measurements) and gradually draw conclusions by refining model/flows/connections/distributions. One day the squares will cease to be totally black and you will be able to trade on the model. And you cannot trade just on any distribution
And we should add something that everybody knows, but forgets about:
- any transaction passes through the market glass twice - at the opening and closing, i.e. there is a "time of transaction holding" distribution.
- The total volume of the exchange can be considered constant, and do not take into account entry/exit volumes. Whoever bought the lot, will sell it later, someone has earned as much as someone else lost :-) So there is a "time to market" distribution
- on the timeframes, there is no price formation, but rather price refinement within given ranges. If the price is close to the limits, the regulators of different scales come into play.
You have to accept it for what it is and, by loving it
Smart thinking comes after the fact.
and also add something that everyone knows but forgets :
- any deal goes through the market glass twice - at opening and closing, i.e. there is also a "deal holding time" distribution.
- The total volume of the exchange can be considered constant, and do not take into account entry/exit volumes. Whoever bought the lot, will sell it later, someone has earned as much as someone else lost :-) So there is a "time to market" distribution
- On timeframes, the price is not formed, but it is refined within given ranges. If the price is close to the limits, the regulators of different scales come into play.
but that's not all :-)
As end customers, we almost always deal with aggregated data. That is, with some complex of two or three or more stacks.
For example, again simplified: the server takes liquidity from two sources and they arrive alternately, in A the bid is higher but the volume is 12, in B it is lower with a volume of 30.
A or B (usually A, for spread narrowing) or even something in between may be selected from the settings. How the stacks will be aggregated is a mystery.
But the ticks will fall with terrible force - the frequency will be higher than at A or B. Their prices are approximately equal (and the connection has become good and algorithms tightened, and arbitrageurs have done their job),
but Bid will vibrate +-point-two, which the especially suspicious take for intentional/ malicious action.
It's not even easy to collect data here :-) if taken from an arbitrary, there are more likely to be clear aggregation effects. If one storms the logic of the cup/market, it is long and tedious to choose a centre with a large and constant own turnover,
and there will be a higher spread, less frequent ticks and tighter conditions.
I have noticed that if, when leaving the variance channel, the coefficient of kurtosis of the increments is >10, then the trend starts without a "return to the mean". Under 10, there is a return. Sitting - checking. As always - one. "By yourself, all by yourself" - understandable, what's to say...
Maybe you should look at this indicator on related currency pairs that have similar symbols? Anything interesting there....
What has physics got to do with it? direct projections of physical laws do not work here. e=mv^2 does not apply here, for lack of e,m,v :-)
I have even given direct instructions before - start with the basics, with economics, with pricing... You can make money in the market, but you have to work in it all the time to do so. (nonsense, yes, who would have thought?) There is no universal "market formula", no "unique distribution", no "pikey dictates"...
I like all the topics raised by you and Oleg, with one small nuance - they "do not cling" to reality, they are as if in themselves, for their own sake, hermetic. It's an amusing mind game.
You looked at the returns almost under a quantum microscope looking for a specific distribution - well there should have been a shadow of a thought, by what process, in what way could this distribution be obtained ?
but it didn't happen... you goddamn mat-theoreticians, pardon the expression.
Stochastic indicators show something for some reason. In every terminal and 10 at different timeframes. Why do people use them? They show some kind of pattern buried in the fastest indicator - the rate.
Stochastic indicators are showing something for some reason. In every terminal and 10 at different timeframes. Why do people use them? Some kind of pattern buried in the fastest indicator, the rate, they show.
Look at what the stochastics are about. It has nothing to do with statistics but shows price position relative to max/min on the interval.