Random Flow Theory and FOREX - page 26
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
What subtle features? For example, the discreteness of the market (Fiboidism in some perverse form), the presence of channels, support/resistance levels.
If you know how the process is organized: generate several dozens of thousands of "synthetics" (in the "worst" case - which actually is the best one for checking the robustness of the TS) and check directly the % of them, on which the TS loses. And then using the ergodicity hypothesis invented especially for the sake of justifying the testing, apply the statistical checking results obtained in the process realization space to the real trading in time.
Yes, it is long, but it is much shorter and cheaper than checking with own money in real time. However, the smaller required statistically reliably detectable % of TC killing (probability of killing - p), the more synthetic is required (statistically sufficient number of synthetic for small p is a function of order 1/p^2). But you don't need any forward analyses, multicurrency and multiperiod tests and other rubbish described by Pardo - but only if you know the model of process, adequate to information used by TC (I think I start to intuitively understand, what sigma-algebras are needed for in tervers :))).
P.S. 2 Prival: I'm reading Amir's article ( 'The principle of substitution of time in intraday trading' ), searching for something to please your inflamed imagination looking for physical analogies. Look:
Why, there is nothing easier - if you know how the process works: generate several tens of thousands of "synthetics" (in the "worst" case - which actually turns out to be the best for checking the robustness of the TS) and directly check what % of them the TS is losing. And then, using the ergodicity hypothesis that was plucked from the finger specifically to justify such testing, we extend the results of statistical testing obtained in the space of process realizations to real trading over time.
I've already written that I want to make the proof of robustness/non-robustness of the TS automatic (statistical, of course).
Won't this thing kill all 100% of TCs ?
Hi Rosh. That's exactly what I'm looking for, i.e. how. The main sticking point is stationarity. I don't actually need Gaussianity, vinoriness, martingale and so on. What I need is the stationarity of at least some transformation of the quoting process. In principle, Amir's article mentioned earlier gives hope for that (for stationarity of equivolume bars). Hardly so on large TFs (like weeks), but it seems true up to 4-hours inclusive. Further on - only technical coding problems, which are not crucial.
Strong mathematicians-mechanicians, who know statistics, are here. Will do research or tell me where to dig, I will be very grateful. Once again: we need to confirm the stationarity of at least some reversible transformation of the quoting process.
About specific pairs... I don't know yet, I've looked only at oira, at least for it, because I suspect that the laws are different for different instruments. We like oira the best anyway, so even in this case there will be a benefit.
Candid, very likely - for TCs based on normal continuous filters. less illusions would be involved.
I've already written that I want to make the proof of the robustness/non-robustness of the TC automatic (statistical, of course).
Will this thing not kill all 100% of TC?
It will kill all of them ! Except those whose authors manage to agree with Mathemat. :-)
Yep. Someone does know a counterexample to a robust system based on, say, traditional trading interpretations of the wagons, RSI, stochastic, alligator, momentum and whatever else is continuous. (By the way, I don't rule out the possibility of such.)
And it will be difficult to agree with me: I changed my avatar to a radical revolutionary one, when I clearly realized, that something serious and constructive can come out of this idea, and not just a criticism of the principles of TS building :).
P.S. 2 Prival: I'm reading Amir's article ( 'The principle of substitution of time in intraday trading' ), searching for something to appease your inflamed imagination which seeks physical analogies. See:
A HUGE THANK YOU GAVE ME WHAT I WANTED, what is most important, what I have been looking for, for many years now since my 1 account went bust. Now it (the market) cannot beat me, even if all the financial analysts in the world sit on the other side of the monitor. Grandpa Elder was right. It's what's in your head that counts!
Therefore, this curve on the screen for me will NEVER be a Brownian motion (martingale, diffusion ...) or whatever they call it, I no longer care about all these names, and the trading system - a game of wagering (and all the theories associated with what you can't beat).
This curved trajectory of the enemy, cunning, insidious and ruthless. And the entrance there is not (buy, sell, bid or whatever it is called), but a rocket launch, and the exit point (taking profit, TP, SL or whatever) I have a probability of defeating the enemy.
Now they can't beat me as they have this game and the stake in this game is ruble. I have a war and the stakes are INCREDIBLY HIGHER and if he (the market) wants to win (= come rape my wife and sell my children into slavery). He must first kill me, dance on my corpse and wipe his feet on me.
And if it's a war (imagine this situation), and on my side of the monitor are real fighters (knights, warriors, soldiers - the military in a word), while on the other side are Buffets, Soros, Ben Ladins, etc... they have NO chance of winning even at 1:10^100 I wouldn't bet a dime on it.
Larry Williams, Batter and others are showing us that it is possible to beat the market.
P.S. Mathemat thanks again, now I understand what I'm doing. Now I have no more "inflamed imagination", only cold and sober calculation.
Edit. It's impossible to defeat me now, but can I win? If I manage to find even the slightest patterns in my opponent's behaviour, I will put them in my piggy bank and try to use them. I'll study it long and hard. And when the probability of defeat reaches the order of 0.7 I will go out to fight.
No you can't take me with your bare hands now, there is no TP/SL and system (5/100), I have to hit the target. 7 hits out of 10 (and a miss is 0, not a minus.). Profit (wounds) we'll count later :-)
Hi Rosh. That's exactly what I'm looking for, i.e. how. The main sticking point is stationarity. I don't actually need Gaussianity, vinoriness, martingale and so on. What I need is the stationarity of at least some transformation of the quoting process. In principle, Amir's article mentioned earlier gives hope for that (for stationarity of equivolume bars). Hardly so on large TFs (like weeks), but up to and including 4-hours it seems true. Further on - only technical coding problems, which are not crucial.