Discussion of article "Extract profit down to the last pip" - page 17
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
MO maps from the trait space to the target traits, i.e. there is no need to prescribe the rules of trade.
If genetics just gives a selection of variants and does not tell anything about quality, MO generalises and can give additional information.
If you formalise the regularity of your TS, it is not difficult to make MO for this purpose.
We can hack your algorithm through MO, but the problem is that it works only in one dts (as far as I understand).
Guys, I'm trying to follow the discussion. Am I understanding "MO" == Machine Learning?
Original: Guys, i try to follow the discussion. Do i understand correctly "MO" == Machine Learning?
Guys, I'm trying to follow the discussion. Am I understanding "MO" == Machine Learning?
Yes.
If you formalise the regularity of your TC, it's not hard to do an MO for that.
We can hack your algorithm via MO, but the problem is that it works only in one dts (as far as I understand)
Train on the inputs in the hope that the resulting MO-TC will be better than the original?
Train on the inputs in the hope that the resulting MO-TC will be better than the original?
maybe it will optimise (learn) faster, there may be different variants, it is not clear in advance
Train on the inputs in the hope that the resulting MO-TC will be better than the original?
Machine learning is done using an algorithm written by humans. The learning results (optimising the input parameters, finding the best combination of input parameters) will be better unless the initial parameters are the best, which may be random, and definitely will not be in most cases of optimisation. Optimisation is a re-selection of results among many choices of combinations of input parameters. In the graphs for three this is the surface, and unfortunately in this kind of optimisation there is no analysis of where the best combination is, on top of a steep mountain or on an elevated plateau. A steep mountain means that a small change in the input parameters of the TC will sharply deteriorate the result of the TC, on a plateau in the middle we will have a stable zone, on the edge of the plateau deterioration will be when changing one of the parameters. Speculatively it will be correlated to the change of price behaviour, on a steep mountain changes in price behaviour will much more often lead to deterioration of the TS result than in the plateau zone.
Forum on trading, automated trading systems and testing trading strategies
Discussion of the article "Scratching out the profit to the last pip"
fxsaber, 2019.09.25 06:26 pm.
Two months.
Three months. The advisor and settings have not changed since the start.
When low mat. expectation breaks the result.
In October, the TS trampled on the place(static and interactive). And this is despite the tolerable stability of scratching out pips(static and interactive).
It cannot cope even with the commission below average on retail. We expect a plum.
Write a sequel. On these points.
Some unexpected figures from the Tester.
It is not a secret that mostly illiterate traders trade marquees. I wondered, what about the TS on limit orders?
Forum on trading, automated trading systems and testing trading strategies
New version of MetaTrader 5 build 2190
fxsaber, 2019.10.31 11:38 am.
So how many modifications (trade orders)?
Real example. Two months, only one position open (hedge) at a time and an order to open the opposite position when the current one closes. Trading only 12 hours a day.
I don't bring extreme figures, where there are a lot of positions, but I take only a good pass, which can be put on real.
Positions - 851.
Modifications - 1 300 513 (more than a million).
~1500 modifications per position.
And these are not some stupid modifications from nothing to do, but working. In real trading there are many times more, because there are a lot of TCs.
Trading servers cope with such a load without problems, so I have not even tried to artificially limit the number of trade orders. There are possibilities for this, but why? Probably, it is only expedient for the exchange.