Topic: Same Direction signal - Winner

What is considered to be a Winner? 1 pip above/below the entry point or 1% gains in position?

Re: Same Direction signal - Winner

Nothing of the above.

A position is a "Winner" when the positions Floating Profit Loss > 0.000005 to the moment of the adding.

A position is a "Loser" when the positions Floating Profit Loss <  -0.000005 to the moment of the adding.

Entry point is not a factor because of the swap charged overnight.
Example:

Open 3 lots long at 1.23000.  Spread 20 points. Swap long 2 points per day per lot.

After 5 days the corrected position price will be:

PosPrice = EntryPrice + Spread + Swap*Days*Lots;
PosPrice = 1.23000 + 0.00020 + 0.00002*5*3;
PosPrice = 1.23020 + 0.00030;
PosPrice = 1.23050;

The position will be called a "Winner" if the current Bid price is greater than 1.23050
and called Loser if:  Bid  < 1.23050

Re: Same Direction signal - Winner

Thanks for the quick reply.

What's the average price in the journal: average of high and low price or average of open and close price?

Re: Same Direction signal - Winner

The average price shown in the journal is the current position price. It includes the position modifications - adding, reducing, reversing... (if any)

Re: Same Direction signal - Winner

I have a specific problem with adding winning positions in FSB that is often performed in an incorrect way.
In FST (both demo and real) it seems to be performed correctly.
The result: In 95% of cases a winning strategy backtested in FSB turns into a zero or sometimes even a losing strategy in FST (= in real trading).
To illustrate this strange behaviour I created a sample strategy that is almost identical with Mr. Popov's "Basic Strategy EURUSD 15min" published on 23rd May.
See the screenshot where I tried to put all relevant information together. The top right chart shows candles and trades made in my broker's platform. The remaining views belong to FSB. Both FSB and FST (= my broker's) data is identical. Here are 2 examples of the strange behaviour:
1. Long black candle at 12:30; in FSB the Short position was added (wrong), while in the real trading not (correct).
2. White candle at 6:30; in FSB the Long position was added (wrong), while in the real trading not (correct).
I registered this strange behaviour also with other indicators, in other timeframes, with other underlyings and with other broker's data.


http://s2.postimg.org/ns957noc5/prtsc_2.jpg

Re: Same Direction signal - Winner

Can you attach your strategy?
I'll try to reproduce that behavior on my data.

Re: Same Direction signal - Winner

Here is the strategy. I use FXCM data.

Post's attachments

111_EURUSD_m15_Fxc_Popov-Forex Strategies-repository-20140523.xml 11.94 kb, 3 downloads since 2014-06-18 

You don't have the permssions to download the attachments of this post.

Re: Same Direction signal - Winner

As I mention on another post on the subject, we found a bug in Winner and Loser adding mode. A bug fixing release will be available tomorrow evening.

Re: Same Direction signal - Winner

Thank you for the confirmation.
Let me still ask you a question that may sound like a bad joke (and that you will most probably answer with "No" for some reason that I am not aware at the moment).
Before I sent you my original bug report, I tested the behaviour of this bug quite thoroughly. I must say there was no randomness in it; vice versa, it behaved incredibly consistently in all possible setups and environments. So that a logical question came to my mind:
Why not to try to implement this "bug" into the FST (Trader) part and make it a new great "feature"?  smile

Re: Same Direction signal - Winner

Why not to try to implement this "bug" into the FST (Trader) part and make it a new great "feature"?

There only one, but HUUUGE problem with this bug. It leads to incorrect backtest. The fault code makes the backtest absolutely wrong. I'll explain it short:

Instead of: Add "Winner" when the positions Floating Profit Loss > 0.000005,
The program does: Add "Winner" when the positions End of Bar Floating Profit Loss > 0.000005;

It actually looks in the future and adds only if the adding will result in a profit. This is an absolutely unacceptable issue and ruins the whole idea of FSB - to provide a correct backteting result even when the user doesn't want smile

Re: Same Direction signal - Winner

Popov wrote:

Why not to try to implement this "bug" into the FST (Trader) part and make it a new great "feature"?

There only one, but HUUUGE problem with this bug. It leads to incorrect backtest. The fault code makes the backtest absolutely wrong. I'll explain it short:

Instead of: Add "Winner" when the positions Floating Profit Loss > 0.000005,
The program does: Add "Winner" when the positions End of Bar Floating Profit Loss > 0.000005;

It actually looks in the future and adds only if the adding will result in a profit. This is an absolutely unacceptable issue and ruins the whole idea of FSB - to provide a correct backteting result even when the user doesn't want smile

Is it possible to add some kind of predictive mechanism (or using FSB mechanism to build up this predictive mechanism) to determine if the  positions End of Bar Floating Profit Loss > 0.000005 is true before adding. Since the outcome is only TRUE or FALSE, if the predictive mechanism is much better than 0.5, then we could have a more profitable strategy, called it Predictive Winner ?

Re: Same Direction signal - Winner

"Predictive Winner", how to do this in the trader? The builder can easily see int he "future", but not the Trader.
If you now an algorithm that can predict the future prices with probability > 50%, you are welcome to share it with us.

What you are asking about  is actually a complete FSB Strategy.

Re: Same Direction signal - Winner

Popov wrote:

As I mention on another post on the subject, we found a bug in Winner and Loser adding mode. A bug fixing release will be available tomorrow evening.

Some time ago Mr. Popov added a EURUSD 15m Strategy to Repository (it is no longer available there). I tested this strategy with FSB Pro and FSB with nice linear results (on historical data). Recent upgrading of FSB Pro to v. Beta 2 changed winning results of this strategy to loosing – presumably because Mr Popov removed some bugs (Winner and Loser option?) in the Pro version. But the winning behavior of this strategy is still present in FSB, probably because of same bug - not yet removed?

Re: Same Direction signal - Winner

I removed the strategy because its nice looking curve was a result of the "Winner" bug.

Today I uploaded a fix to that bug for FSB also.