From theory to practice. Part 2 - page 56
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
Logically, that's exactly what it is. But I'm not going to argue about it...
Logic is good... but if it is supported by examples...
The website is full of examples showing reports with stable profits over the long term... These examples clearly do not clash with this postulate...
In any case, all 'logical' constructions require careful checking, however 'logical' they may seem at first sight...Hmm... So, it is impossible to make money on SB by definition
Yes, by definition, the expectation of equity is always equal to zero. In this case, equity, unlike the price, will not be SB, which sometimes leads to apparent paradoxes (like the equity of martingale).
and on market VR - similarly due to non-stationarity. Am I getting this right?
It is more complicated here. On SB it is always zero on average (not including the spread). Because of non-stationarity - a good plus when correctly guessing the direction becomes a good minus when correctly guessing the direction.)
If we bring VR to the stationary form, will anything change?
Trading will always follow the original series) Unless you try to build a stationary spread or a portfolio of several instruments.
In general, in theory, it is impossible to make money. But, in practice, it is not!
The probability of a single event triggering and a total event are completely different things.
Yes, by the definition of SB, the expectation of equity is always zero. In this case, equity, unlike price, will no longer be SB, which sometimes leads to seeming paradoxes (like martingale equity).
It is more complicated here. On SB, it is always zero on average (not including the spread). Because of non-stationarity - a good plus when correctly guessing the direction becomes a good minus when incorrectly guessing the direction) Opportunity to profit is also a potential destructive loss) On the example of your system - sometimes it turns out that you need to trade on the trend).
Trading will always go on a baseline) Unless you can try to build a stationary spread or a portfolio of several instruments.
By the way, exactly the same thing I told Comrade A_K two years ago)
"A good repeat and another repeat!" )
Equity, unlike price, will no longer be SB, which sometimes leads to seeming paradoxes (like equity martingales).
Good.
However, back to the important phrase
It turns out that it is non-stationarity that is the key to unlocking the market.
Non-stationarity manifests itself in both non-stationary intensity of tick flow and non-stationary distribution of increments. And there is nothing to be done about it, for immediately we move to SB.
You have to take these two things "as is" and use them in your TS.
Hmm... There's certainly something to that...
Alexander, well, these are all general concepts. Usually they talk about some specific market inefficiencies that need to be found. They won't necessarily lie only in the plane of one time series. Situation by situation.
It's like taking SB and adding some cycles to it, that's all. You can't fit it to a specific market formula, it's individual everywhere.
As the esteemed Prado here has written
" Key difference between (classical) statistics and machine learning: how they approach the bias-variance tradeoff dilemma. Generally speaking, if you know the data-generating process for certain, use statistics. If not (e.g., finance), use machine learning."Alexander, these are all general concepts. They usually talk about specific market inefficiencies that need to be found.
If we start from a hypothesis:
Forum on trading, automated trading systems and trading strategies testing
From Theory to Practice. Part 2
Alexander_K2, 2021.04.18 10:55
It turns out that non-stationarity is the key to unraveling the market.
Non-stationarity shows itself in non-stationary intensity of tick flow as well as in non-stationary distribution of increments. And there is nothing to be done about it, for immediately we move to SB.
You have to take these two things "as is" and use them in your TS.
it is obvious that
- the unsteady intensity of the flux manifests itself at different hours within a day. You have had research in this area, but no practical results. I may have missed something...
- Unsteady incremental intensity manifests itself in very large impulse motions. I don't think there's been any research on this at all...
And yet these are the most obvious market inefficiencies.
Assuming the hypothesis:
then it is obvious that
- the unsteady flux intensity manifests itself at different hours within a day. You have had research in this area, but no practical results. I may have missed something...
- Unsteady incremental intensity manifests itself in very large impulse motions. I don't think there's been any research on this at all...
And those are the most obvious market inefficiencies.
Looks like the sample size should be changed depending on the time of day. Statistically, the sample size should be increased at hours when you encounter trends.
I wanted to optimize in the mt4 tester, but it doesn't even want to run with 24 parameters.
And why is that? The sum of SB chunks should seemingly give SB (if a constant lot). Which means that it is impossible to make money.
With a constant lot, this is the case. But all attempts to "earn" on SB consist in constant changes of volume and direction of deals. This may lead to complex and bizarre volatility fluctuations and similarity with SB may be lost, although zero expected payoff will not disappear.