Optimisation vs Fitting in the story - page 2
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
Yes, question about in sample. And how is it calculated, accurasi? It's not going to work here. Let's assume there are no parameters at all, let's say we want to fit an optimal SMA.
Why can't the mnc work? you can measure it too. in what parrots you measure that metric will work, the variance is basically
akurasi = number of correctly predicted observations/number of observations
Optimization must be performed in one pass, in the body of the Expert Advisor at the first start before the start of trading, and then during the trading process onlyoptimized parameters should be corrected. The optimization is performed by a special internal separate block of the Expert Advisor calculating the parameters and not by searching through them. In this case, the parameters can be made private and visible only to the Expert Advisor, because they are self-adjusting (self-optimizing).
If there is overshooting of parameters in multiple passes regardless of the use of forward - it is a fitting.
why won't the mnc fit?
Because for a non-parametric curve, say SMA, it won't have an optimum, the sum of squares of deviation will keep decreasing as the fit gets stronger.
How can it be used to trade on the 1st wave? it's just a smoothing with no prognostic features
i.e. what are we optimising in general?Is there an example of this block, in CodeBase?
And if it were, how would it help you?
And if it were, how would it help you?
How can it be used to trade on the 1st wave? it's just a smoothing with no prognostic features
I.e. what are we optimizing?I just want to fix the muving to go like in the middle picture, I can't figure out the metric for that.
The metric can be taken as the number of normal or relevant trades that cover the spread, etc. I.e. total profit, for example, divided by the std of prices from the machine
is a metric at a glance. It is clear that the ISC will not say anything by itself
I mean, if it is a minus-reverse strategy, the rest is analogous
the second option is generalised models, i.e. we take the average of MnC among many models
In my opinion, the standard Mind Sample EA is a self-optimizing Expert Advisor because it has SL written in the condition. If you prescribe TR as well, it will be 100%
It doesn't have a stoploss. There is a trailing function, it puts a stoploss. But it may not reach trailing stops, the order may go straight to the loss, in this case there is a market closing according to the terms of the indicator.
Suppose we can make Stop Loss and Take Profit proportional to ATR or STD, or whatever - here we can say that the parameter is self-optimizing, but not the entire Expert Advisor. There may still be a desire to optimise the coefficients used to calculate stoploss and takeprofit. There is always something that can be optimized.
The interest here is directed towards a self-optimizing Expert Advisor. One that does by itself what everybody usually does in the tester. But you canoptimize the self-optimizingparameters... then make them automatically optimized... and so on endlessly.
Nikolai offers something from the realm of science fiction - not to optimize by searching for parameters, but to calculate a great bunch of parameters, including stop loss, take profit, time limit periods, etc. This task is from the realm of fantasy, not realistic at all.
It doesn't have a stoploss. There is a trailing function, it puts a stoploss. But it may not reach trailing stops, the order may go straight to the loss, in this case there is a market closing according to the terms of the indicator.
Suppose we can make Stop Loss and Take Profit proportional to ATR or STD, or whatever - here we can say that the parameter is self-optimizing, but not the entire Expert Advisor. There may still be a desire to optimise the coefficients used to calculate stoploss and takeprofit. There is always something that can be optimized.
The interest here is directed towards a self-optimizing Expert Advisor. One that does by itself what everybody usually does in the tester. But you canoptimize the self-optimizingparameters... then make them automatically optimized... and so on endlessly.
Nikolai offers something from the realm of science fiction - not to optimize by searching through the parameters, but to calculate an awesome pile of parameters, including stop loss, take profit, time limit periods, etc. This task is from the realm of fantasy, not realistic at all.