From theory to practice. Part 2 - page 173
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
Speaking of the factory, a lot of people here are from there:
paint factory workers also come home through the magical forest....
Probably not just gnomes, but oompah-loompahs from Willy Wonka's factory)
Shiryaev said that if we add an epsilon "outwards" to the repeated logarithm curve, the crossing will happen only a finite number of times, which is of course very good, but in practice it is not clear exactly how many times, and the worst thing is that the distribution is constantly changing, Therefore estimating the iterated logarithm in the low-volatile area may cause a lot of unpleasant emotions when the volatility rises, and you can even wake up in a brick factory, similarly the opposite, volatility overestimation will be harmful for trading, and the iterated logarithm has an undefined value in the zero point and also a shift of the chart by one, And it is easier to estimate the cocoon in any case using the empirical root plus a small linear trend, because in any case the parameters will be lost on a forward position and there is not much sense to build a super-duper accurate model like Shiryaev's one using calculus of variations (about the best price estimation within a day), when there are elections or all sorts of OPEC rips or the like, that's what the foreman told me, well that's it, it's time for me to unload the new supply, I'm off...
Well, some particularly laughing foremen generally argue that in proper price models, price increments have no variance or even expectation (Levy flights) at all. Consequently, the CBAs with DTTs and other such things go out the window. They say there is a conspiracy by brokers and financial analysts to always treat such models as marginal (because they lead to a sharp reduction in volumes).
Well, some particularly funny brigadiers argue that in correct price models price increments have no variance or even expectation (Levy flights) at all. Consequently, the CBAs with DTTs and other such things go out the window. They say there is a conspiracy of brokers and financial analysts to always perceive such models as marginal (because they lead to a sharp decrease in volumes).
I guess the conspiracy in the industry is not to tell a real risk assessment to a wide range of investors 😁
and yet it is possible to talk about some expected (implied) volatility in an ordinary non-election non-covid non-geo-politicised market
I suppose the conspiracy in the industry is not to tell a real risk assessment to a wide range of investors 😁
And yet it is possible to talk about some expected (implied) volatility in the context of the usual non-election non-covid non-geo-politicised market
You can and should talk about volatility - rather, you can't even avoid talking about it.) Although, reading some forums can sometimes give the impression that volatility does not exist and that prices move exclusively along deterministic trajectories, for which you just need to find the right secret formulas).
The problem with Levy flights is that the measure of volatility cannot be variance and the measure of dependence cannot be correlation (due to the non-existence of these quantities).
Volatility can and should be talked about - rather one cannot even not talk about it) Although, reading some forums can sometimes give the impression that volatility does not exist and prices move exclusively along deterministic trajectories for which one just needs to find the right secret formulas).
The problem with Levy flights is that the measure of volatility cannot be variance and the measure of dependence cannot be correlation (due to the non-existence of these quantities).
Of course we should use it, because there is nothing else (of course one can use his own metric instead of stdev, but the essence is the same) and exploit the main repetitive behaviour inherent to a given market and given instruments without violating something that was determined with a high enough empirical probability earlier and taking into account the maximum expected speed (if we do not take any exotics like ruble and lira) for it is the repeating property that system trading should be based on (as opposed to naive discretionary trading), and because
It is thought that the problem with any system with a complex algorithm and a not-so-complex algorithm is overlearning. We look at the price movement parameter on the history, and find the median, the most frequent value. Then we adapt the system to it, run it and... The median value of the parameter will also start randomly fluctuating within certain limits. And our resulting miracles of history are not repeated in the real world.
We judge an infinite population, which we cannot see, by a finite sample, which is one slice of the diversity of that entire population.
Of course we should use it, because there is nothing else (of course one can use his own metric instead of stdev, but the essence is the same) and exploit the main repetitive behaviour inherent in a given market and given instruments without violating something that was defined with a high enough empirical probability earlier and taking into account the maximum expected speed (if we do not take any exotics like ruble and lira) for it is the repeating property that system trading should be based on (as opposed to naive discretionary trading), and because
I think let's not make the same drunken mistakes that you and I would get banned for the joy and ridicule of others)))
Every financial instrument, just like every individual in oYounger society, is individual and at the same time subject to certain laws.
The laws of financial markets are quite clearly defined by nature, but are not available to everyone. After some time, the market will be liquidated, under one common task. That is the management of society. Whether it will succeed without a cataclysm is doubtful. Cataclysm is already a known way into the caveman image for the new advent of fat life awareness.
It is thought that the problem with any system with a complex algorithm and not so complex is overlearning. We look at the price movement parameter on the history, and find the median, the most frequently repeated value. Then we adapt the system to it, run it and... The median value of the parameter will also start to fluctuate randomly within certain limits. And the former miracles of history do not repeat themselves on the real.
We judge an infinite population, which we cannot see, by a finite sample, which is one variant of the diversity of that entire population.
Overlearning.
Sorry to barge in without asking.
Overlearning can only be for certain stretches of time. So I understand.
But if the time has changed, retraining will already be required. And so there must be no interruption inretraining.
Well, I certainly envy such TCs.)))
Overlearning is in the sense of adjusting to history. We try to learn from history, and tomorrow, or the day after, or the week after, what we have relied on floats on its own. Overlearned means.
Overlearning is in the sense of adjusting to history. We try to learn from history, and tomorrow, or the day after, or the week after, what we have relied on floats on its own. Over-educated, then.
Yes, retrained, or even found a fake pattern, an illusory, false, maya of chaotic dancing shadows.