Market etiquette or good manners in a minefield - page 96
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
No, no, clusters are not martingale at all, it's different. That's exactly what I'm talking about. You have to look where the market is not martingale. But pair quotes (without information from other pairs) are almost martingale.
+1
The second half of the answer: they are generated by the server module's MT-4 ticogenerator (with DC settings, of course).
So the process is actually mechanical.
Can you please give me a link to MT server datasheet or any other source documenting mechanical/artificial origin of tics in MT.
Fighting with currency portfolios. But I am not going to talk about them here, I am also being greedy. In fact, all that's left from Semenych is the basic idea - that there are clusters, so they should be traded. Oh man, when will you colleagues stop discussing this damn near-martingale, hoping that the output will be non-martingale... Clusters are not martingale at all, it's different. What you should be looking for is exactly where the market is non-martingale. But the quotes of the pair (without information from other pairs) are almost martingale.
Alexey, don't be greedy, put your hypothesis - let's discuss it. After all, greed is a sin! Why do you need it? So you may get some wise advices, and we'll save you several years of idle hard work on the hopeless idea... Well, I'm only joking. Seriously, from my experience on the subject, I can say that in the end, it all comes down to exploitation of short-lived arbitrage opportunities. This is when a sharp movement detected on one or several instruments provokes a statistically significant movement on other instruments. The effect is real, but (as always, there is a BUT) the time in which the market becomes effective and destroys the arbitrage, much less than the characteristic time of the order execution by the brokerage company!
Alexey, I don't believe you have found a way around this ban! Or do you have a dedicated trading analogue?
The fact that the market quote is not a martingale, you don't deny it yourself. The question is the absolute value... Yes, not much. Yes, almost everything is at the level of DC commissions, but it's the BEST and there are degrees of freedom (NS) to further optimise the strategy.
On the subject of martingale:
Below is a quote from http://monetarism.ru/article.pl?sid=06/05/21/1328256&mode=nocomment
"...Martingales are closely related to the notion of an efficient market. Weak efficiency follows from price martingale. However, medium and strong efficiency does not follow from weak martingale and there is hope for a profitable strategy with additional information(not contained in the price history).
Market makers (central banks for currencies, stock exchange specialists for equities) use martingales to generate prices. This is very convenient, combats unwanted feedback, provides high liquidity and creates economic stability. On the other hand, high volatility martingales can be perceived as undesirable speculation.
Even if you have statistically proven that a series of prices does not belong to the martingale class - rejoice too early. If a market maker has a martingale that consists of numbers consisting of exactly 3 decimal places (with an accuracy of 10 pips), the market maker (or broker) can add a 4th decimal place to the theme in a special way, creating asymmetry and obtaining a series that does not belong to the martingale class. However, due to the existence of a 5 pip spread, we will not be able to obtain a profitable strategy, as all of the non-martingale sits in the 4th sign, which is eaten by the spread and reduces profits.
The first evidence of the martingality of stock prices is present in Bachelier's dissertation of 1905.
Despite the 100 year history of martingale theory, most people still believe that there are trends, cycles and flats in stock prices... "
Apparently there is the question of the method of calculation and the amount of fluctuation (H for kagi) to get out of martingale
Apparently there is a question about the method of calculation and the amount of oscillation (H for kagi) to get out of martingale
Alternatively:
Evaluate the predictability of p RT in one way or another (NS, classifier, etc.) on different H. Obtain the dependence p(H) . Choose a North where p is maximal. Work on this trading horizon and in the background scanning the whole H range, until the mood changes (the North changes ).
Alternatively:
Evaluate the predictability of the p p RT in one way or another (NS, classifier, etc.) at different H. We obtain the dependence p(H) . Choose a North where p is maximal. Work on this trading horizon and in the background scanning the whole H range, until the mood changes (the North changes ).
Perhaps simpler: RMS, RMS, variance, etc.?
But in both cases the question of the reliability of the estimate - i.e. the boundary values of the estimated parameters - will arise.
What ceiling are they written on?
But in both cases the question of the reliability of the estimate will arise.
Obviously, in this formulation the question remains unanswered, because in the case of weakly stationary processes, the uncertainty principle applies - the more precisely we try to estimate a quantity by increasing the statistics, the more likely it is that the "true" value of the quantity itself will change over the statistical collection period, and conversely, the more crudely we estimate the "true" value, the less will the estimated quantity itself change. There is thus a limit to the accuracy of the estimate, which is independent of the size of the statistical sample.
Of course, this does not mean that you can do anything. Simply, one should search the literature for estimation criteria for quasi-stationary quantities.
No, no, clusters are not martingale at all, it's different. That's exactly what I'm talking about. You have to look where the market is not martingale. But pair quotes (without information from other pairs) are almost martingale.
What is not martingale in clusters?