WACKENA unravels 2007 championship deal analysis - page 3
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
I've had this one since 1 January 2007 until now...
Can you tell me what "general" methods can be used to make this graph smoother by sacrificing profits?
I have this one from 1 January 2007 until now...
Another WACKENA:)
YuraZ, stupid poll - where's the code. i couldn't find it?
This question has already been answered:
William Boatwright. WACKENA took second place, he sells his product supports users created website... he's got business going on.
and i'm gonna put out the code... why? even if it's not WACKENA code but it's close to it...
out of respect for him I won't do it.
----
p.s.
the code is almost intimate!
----
the discussion is on the level that CHAMPIONSHIP provides rich fodder by analyzing leaders' stats
looking at logs and some of the subtle information flows, communicating directly with the authors, gives a lot of ideas...
the discussion is about what optimization is and how to do it properly and according to what criteria
2-in an optimized sample I choose for example a set that gives or the same profit let's say
I have 2000-3000 thousand rows after optimization in the optimization results window
let's say we have 70 lines of sample give the same profit and the same drawdown the next sample of 50 lines a little worse profit and drawdown
For example, SL or TP greater than a certain size simply will not be reached.
In my opinion, it is not correct to take parameters from this area. Although, it depends on the parameters, and on the strategy.
Andrew!
That's not what I mean! If you optimize a single parameter, yes...
And if you optimize a group of them at once, it's a bit different... You don't choose a sample from the mass of runs where dancing on parameters doesn't change anything
but a sample that gives approximately the same results
in this sample, stable parameters with small variation are drawn!
i.e. something in between...
and one more thing! one subtlety during optimization
1) You should choose 2-3 months of the trend and optimize only on it
2) You should choose 2-3 months of falling trend and use it for optimization
3) We choose a flat period
we get 3 groups of parameters and it is possible to switch this or that set on and off manually once a month or once every 3
turning days are often quite visible
you get 3 groups of parameters and you can switch a set on and off manually once a month or on 3 headland days are often quite visible
Here I will also admit that I do not agree completely.
First, trends may be of different strength, which may affect the performance of Expert Advisors.
Secondly, trends may be accompanied by high or low volatility. In other words, it is one thing when the price goes down smoothly, and it is another thing when the fall is accompanied by 40-50 points candles.
Thirdly, as we can see in January - the trend direction can change after the news release. For example, on the EuroFount currency pair, on January 15 the trend turned sharply down, and yesterday (February 01) it sharply turned up.
Sorry for the criticism - but I am in favour of long term testing, at least for the last year.
you get 3 groups of parameters and you can switch this or that set on and off manually once a month or on 3 turning days are often quite visible
the Expert Advisor should enter - exit - while I'm away from the car... I'm at work - sleeping, taking a walk, etc.
and do it with a high probability of success and preferably by trend
If I'm at the car, then the Expert Advisor set with the priority to go by trend, so it's smart to reverse on this day - with the priority to buy.
reversal day is not always evident, of course, it is a subjective evaluation, since 22.01.2008 I only bought with my hands... almost 2 weeks
+ the Fed cut the naira...broke the previous high... There were 4 points in total with buy signals but i am at work now ...
i posted the picture above ...
best regards
I don't completely agree with this point either.
First, the trends may be of different strength and it may affect the Expert Advisor's performance.
Secondly, trends may be accompanied by high or low volatility. In other words, it is one thing when the price goes down smoothly, and another thing when the fall is accompanied by 40-50 point candlesticks.
Third, as we can see in January, the trend direction can change after the news release. As an example, on the EUR/USD the trend went sharply downwards on January 15, and yesterday (February 01) it sharply turned upwards.
Sorry for the criticism - but I'm a supporter of long term testing, at least for the last year.
the annual testing will probably give smoother parameters, for example in the preparation for the Championship the optimisation was in the period 01.01.2007 - 19.01.2007
I'm talking about training of parameters for sell and buy in short term trends ( in terms of medium term ) 1-2-3 months
I'm talking about methods to choose among a variety of parameters that have the highest probability
profit
If we take the Championship period, the trend is actually UP, and only at the end of it, 23.11.2007 there was a reversal.
if you analyze the Expert Advisor BETTER, it rushed forward only around November 12
and there is nothing surprising the system is trained to buy... the system would not have gone up if it was run on 23.11.2007
a reversal day is a week in principle visible .... at least i managed to see it
it would be nice to reverse the buy settings on that day and set the sell settings
it is more likely that Friday 01.01.2008 was a reversal day
The euro passed through 1.4966 - 1.4921 - 1.4935 for the 3rd time ... low flat on the D1
I managed to almost exactly repeat WACKENA's Expert Advisor, which took 2nd place...
...on the range where the optimization was done there are lots of losers (I'll post them later)
WACKENA ran his EA on a daily chart. Do you think he really took data from the fifteen-minute chart?