Machine learning in trading: theory, models, practice and algo-trading - page 1688
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
OK, some of the answers already give at least something in the direction of searching for information
well, again to the same question .... why part of TS may pass forward testing, and part of TS does not pass - we need a comprehensive assessment of statistical indicators, though.... I suspect that now we will have to unload statistics for each TS and on the basis of statistical indicators make a choice of TS and optimization - and see what will come out - i.e. another "rule of thumb")))
Can you call it the poke method? It's kind of like analyzing results by parameters....
Do you call it the poke method? It's kind of like analyzing results by parameters....
Yes, it's a rule of thumb.
There are established (hard to guess who) rules for TS selection - there should be the highest MO, the best FS, the best continuous profit / loss
i'll have to "mark" TCs that have passed optimization with data on their statistical indices and optimize them on forward - it's a rule of thumb, but what if i see the optimum combination of statistical indices... taptology, however
i don't see another way, and i've come to this in the discussion, video from@Aleksey Nikolayev about Savvateev ))) - about this
Yes, it's a rule of thumb.
There are established (hard to guess who) rules for the selection of the TS - there should be the highest MO, the best FS, the best continuous profits/losses.
i'll have to "mark" TCs that have passed optimization with data on their statistical indices and optimize them on forward - it's a rule of thumb, but what if i see the optimum combination of statistical indices... taptology, however
i don't see another way, and i've come to this in the discussion, video from@Aleksey Nikolayev about Savvateev ))) - about this
I agree that this is a rule of thumb. We do not know what to look for. But just a dumb search by parameter matrix with learning and switching from level to level of GA or neural network algorithm speeds up the process. Doesn't make it more accurate.... but somehow it works))) There is no logic in evolution. It is an axiom.
I agree that this is a gauge method. We do not know what to look for. But just a dumb search by matrix of parameters with learning and transition from level to level of GA or neural network algorithm speeds up the process. Doesn't make it more accurate.... but somehow it works))) There is no logic in evolution. Axiom.
Again, the first time is not accurate. The goal is to find what to look for. Significant parameters.
There is some intuitive feeling that this model can only give something between SB and flat.
To describe trends we need some other game model. Perhaps we should ask Savvateev.)
In any case, it is unlikely to get a model that gives trends and flies and transitions between them (as it always is in reality).
I had a look through Savvateev's YouTube channels; although he is a positive man, the Internet community can make anyone lose his temper.
Why do we turn off comments (and in other cases, we mercilessly ban them)?
Because to comment come medical morons who can not categorically answer anything on the merits of my arguments, wailing that eh so sorry that such a brain is so crazy. If there's one thing that really pisses me off, it's people like that. Even more infuriating are scumbags who add things like, "Yeah, he's already got it wrong in math," without giving any examples, of course. Everyone makes mistakes, they all get better, and that's fine; when some jerk accuses me of losing my brain in math too, frankly I want to quit and join the snipers :-)) Unfortunately, normal people, most of whom are on our channel, don't think it's necessary to put the boor in his place - people are too lazy, or don't have time, or don't want to feel dirty about the boor - and then all that is left for us is to ban and/or cut comments. This has nothing to do with constructive criticism, but there is less and less of it (and where to get it, if I don't get involved in those matters that I am not familiar with).
I think it is possible to be misunderstood by him in these troubled times with your incomprehensible questions (((
Keep watching the video
When smart people talk to each other, and you can only understand that you don't understand a thing.
And you want to say something clever in response. You know, something to make a point...
But except for YA CREVEDKO nothing comes out. Ehhh.
People here think that getting money out of a trader is a difficult task.
so they're trying to use artificial intelligence
But to go up against a big risk so that the risk increases to a stop-out is an elementary task.
that's why you can't apply the MO, no matter how nice the tests look
hmm... hard, I'll try again, but once again: there is no task to search for TS, no task to determine the trend-flat
1. there is a set of strategies which showed good results at the test
2. there is a subset of this set of strategies which showed good results on the forward
3. there is a statistical estimation of the strategy tester
what is the difference between pp. between pp. 1 and 2 ?
is it possible to analyze items 3 and find differences between items 1 and 2 ?
how to evaluate pp1 and pp2 from the point of view of .... what the hell is the difference between them ? - how do they differ?
Unfortunately, practice has shown that simple optimization, no matter the gene grid and so on, gives little usefulness at all. It is almost pure fitting. When they say "good result on a forward order", as a rule it's an accident, because there are not many deals (less than a couple hundred), and even with redistributive risk management, like martin, taking the loss "out of brackets (forward period)". So you can make a pretty piece of shit and sell to the naive, but not in any case trade by yourself.
From the statistical criteria, I would suggest an obviously high anual (do not remove the letter y) Sharpe ratio on the forward, such as above 3, if the transactions are at least more than 200, and preferably thousands. This is all provided that the trading infrastructure without errors, no peeping, etc. Well, of course this is without MM, trading with a constant lot.As a rule, it is either random, because there are few trades (less than a couple hundred)
adjust optimization by user criteria
trading strategies of interest to me are all 500+ trades within 18 months of optimization (forward is another 6 months)
and with redistributive risk management, such as martin, which removes the loss "over the brackets (forward period)
set up the optimization by user criteria
set the maximum risk, above which it does not make sense to complete the test run
Well, I have a feeling that the "accepted" form of risk management - risk in % of the deposit is a straight path to deposit draining, because TS has a short time after optimization, then, in most cases, the parameters of TS don't coincide with the market and the logical saturation of equity occurs ... and here we are getting the rest of the deposit with our % of equity
I.e. most probably it makes sense to use the value of % of the deposit taking into consideration performed deals but with inverse dependence.
I scrolled through Savvateev's YouTube channels, although he is a positive person, but the Internet community can make anyone lose his temper
I think you may be misunderstood by him in these troubled times with your incomprehensible questions (((
let's keep watching the video
I'm just kidding) I'm afraid that I won't even be able to clearly formulate my question).