To use chat, please login.
Back to contacts

NightRide v2.26 - EURUSD (By bcsendes)

The user has deleted this system.

NightRide v2.26 - EURUSD Discussion

rivy
Jul 09 2010 at 03:56
4 posts
I would like rent a signal,please send me the link, thanks .

linuxpower (linuxpower)
Jul 10 2010 at 13:45
20 posts
zizinvest
Sep 19 2010 at 18:04
8 posts
bcsendes
Sep 24 2010 at 00:50
13 posts

zizinvest posted:
    Send me more info


I've just sent you some more info in PM.


Balázs

stephanusR (stephanusR)
Sep 29 2010 at 14:07
216 posts
bcsendes posted:
    
stephanusR posted:
    Back tested systems don't work they suffer from data-snooping bias. look it up on wikipedia.


Stephanus,

I need to disagree with you or if you like I just partially agree with you. If you carefully read this article
it says that the 'MISUSE of data mining techniques that can lead to bogus results'. In our case this is the backtesting.
I totally agree with it, which means that if you write an EA with a certain number of parameters (more is better),
and start to backtest it heavily, I'm sure you will find some setups that results a positive balance. In this case you
hardcode / hardwire the past, hoping that the future will be the same. And that's a failure. A misuse of backtesting.

But I believe that if you want to develop a successful EA, it MUST preform on backtesting. So it's a required criteria.

So the key is the right usage of the backtesting tool. I think my quite sophisticated development / testing framework
satisfy this criteria. But time will tell...

Balázs


The strategies section is only backtested , you have not yet forward tested this system. Unless you have an account under https://www.myfxbook.com/members/bcsendes ? With back tests we can run any set of arbitrary parameters , trying out thousands of different combinations. But this set of parameters will not work in a forward test because history never repeats itself.

Your back test ended May, if you were to back test from June, July, Aug, Sept then the system will fail.

Note that I differentiate between history and repeating patterns. It is possible to time the markets but this is based on patterns within historical events. These historical events EAs can't cope with

stevetrade (stevetrade)
Sep 29 2010 at 14:23
1408 posts
Stephanus,
I guess this is his live system here
https://www.myfxbook.com/members/bcsendes/nightride-40/33066

EurUsd currently 41% up since end of May this year

So, looks like it isn't failing

11:15, restate my assumptions: 1. Mathematics is the language of nature. 2. Everything around us can be represented and understood through numbers. 3. If you graph these numbers, patterns emerge. Therefore: There are patterns everywhere in nature.
stephanusR (stephanusR)
Sep 29 2010 at 14:29
216 posts
stevetrade posted:
    Stephanus,
I guess this is his live system here
https://www.myfxbook.com/members/bcsendes/nightride-40/33066

EurUsd currently 41% up since end of May this year

So, looks like it isn't failing


Thanks for pointing it out. We don't know his strategy thus we don't know whether the account he labeled NightRide uses the same methodology as this one. Note that saying he does use it is not acceptable , we need proof. But since he won' t reveal his method we can't falsify his assertion. It might be true or not, we can't say. Also note that the account had two months of negative equity. Just one month of negative equity would be a red flag.

stevetrade (stevetrade)
Sep 29 2010 at 14:36
1408 posts
True.
You know I'm not even sure that existence is real anymore. Am I a man dreaming I'm a butterfly or a butterfly dreaming I'm a man. Who knows.

11:15, restate my assumptions: 1. Mathematics is the language of nature. 2. Everything around us can be represented and understood through numbers. 3. If you graph these numbers, patterns emerge. Therefore: There are patterns everywhere in nature.
Please login to comment .