All Things Statistical - 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
"You can look for patterns where, on average, all the models out-of-sample continue to do well. You know you are doing well if the average for the out-of-sample models is a significant percentage of the in-sample score. Generally speaking, you are really getting somewhere if the out-of-sample results are more than 50 percent of the in-sample"
http://stats.stackexchange.com/questions/31513/new-revolutionary-way-of-data-mining
I think the point stated there is that over-fitting the model can also result in big profits OOS, so be wary of just looking at profit numbers obsessively.
Yes, that is a good point. It could also be that positive out of sample results could still be due to randomness. Trying to use the past to predict future outcomes is always full of surprises. Using correlated assets is one method I am using frequently. It is why the above model is built on the dollar index and not EURUSD or GPBUSD directly is to specifically avoid over fitting.
EMD Vs SSA Vs Kalman
Been looking lately at EMD, SSA & Kalman Filter to see which is better and the relevant assumptions involved.
EMD: signals spread over 2 or more intrinsic mode functions
http://tec.earth.sinica.edu.tw/research/report/paper/20070711HHT.pdf
SSA: Complex trends + inability to differentiate 2 components close in frequency + non-casuality requires sufficient length to reduce continuous recalculation over every bar.
As stated from here:
Where’s the Magic? (EMD and SSA in R) | (R news & tutorials)
Kalman: supposedly some say no assumptions about noise having to be of Gaussian nature while others dispute the contrary. + Linear filter.
If only there was a MT4 indicator to record values down for the 1st two indicators, would be nice if someone could post it lolz (dont like endpointed SSA personally)
Well, anyone have any opinions feel free to comment
Martingale Strategy
A very interesting Martingale strategy that went on for 3 years live and grew 125 times
with over 5200 trades. Which seems now to crash almost, sadly. William Eckhardt once said that the high kurtosis features + the non-stationary aspects of the markets equate to strategies reliability being uncertain until several thousand trades. Positive Expectancy is a difficult issue indeed
Martingaling - Looks so good until the inevitable @ Forex Factory
Wintersky
The usual coin toss signal selling : start 50 signals. One will work - for a while. Usual scamm
Almost all signal sellers at mq are doing the same : they register multiple demo accounts with metatrader, their account type was shown as "real" (irt was like that till someone noticed that metardaer accounts always show "real" even when demo, and them metatrader simply removed the "real" from the description of the signal) and then they start the same thing with slightly different parameters or using different symbols (using different user names of course)
That "signal" market is flooded with signals like that. No wonder that things like that martingale case are happening : those are simple demo account gammes sold to people that were convinced (by metatrader among others) that those are real accounts and that the "authors" care of their own accounts. Just another scamm helped by the one that should not be doing that at all
Real Account
According to that guy a forexfactory, that was a real account.... went from about 1k USD to 125k USD before climaxing....
There are multiple real-life Martingale stories like that no doubt.... highly likely due to the problem with Money Management being faulty. Moral of story here perhaps is that System Positive Expectancy is still the most crucial.
I have a simple question to all : if an account that anybody has, grows from $1.000 to $125.000, what would be the reason not to withdraw any money from it?
Occam's razor principle suggests simplest solution for that
I have a simple question to all : if an account that anybody has, grows from $1.000 to $125.000, what would be the reason not to withdraw any money from it? Occam's razor principle suggests simplest solution for that
no withdrawal - EVER!...because $125K is simply NOT ENOUGH!
What signal was that?
I am asking that because metatrader 4 signals are there for less than a year, and for metatrader 5 (that is there for longer) signals maximum number of subscribers is 16 subscribers for a signal that is not a free signal - majority is having 0 subscribers if they are payed signals