Improving the Losses Percentage

[Gabytrader]

Could anyone supplying me with many Bad Trades Charts(there must be also at least 10 bars after the stopping-out event )!
This will help me to write out a method/ a statistical method to reduce the ammount of losses! i'll share it to all of you!
Thanks in advance to Tony !
Gabytrader

[Matt Bowen]

Hi Gabytrader,

I have included a link to all of the trades taken under the RT Roundup Index trades for all of January. Becasue of the file size I will make this link available for 1-week. After that the message will be deleted.

You will have plent of trades to do your analysis and this will give your more than 30 for a sample size.

This will help me to write out a method/ a statistical method to reduce the ammount of losses! i'll share it to all of you!

Are you trying to fix the Win/loss ratio of the system? Manage the money, not the analysis... ;)

[Gabytrader]

Hi, Tony!
I analyzed all the 31 bad trades!
It takes me all the reat of 190 in order to check the uniformity( month-by-month) of my results.
Here are the results:
31 initial bad trades ( since middle/end december 2004 till today)!Means
-31 "bars"
One stage filtering: -19 bars lost only
Two stage filtering : -10 bars lost only
Special filtering : -6 bars only
The filtering method didn't look back in the past, it is simple to put in practise and don't risk more than the same initial risk of the first Enter!
Moreover, it has a good reason to work good! It take account to the statistical mechanism by means of which it must be also bad trades!
Let me have please, all the rest of 190 bad trades!
Gabytrader

[stevej]

Gaby,

This will help me to write out a method/ a statistical method to reduce the ammount of losses! i'll share it to all of you!

Be interested to know what filters you are testing.

How many good trades are filtered out with the bad?

What is the win / loss rate using the filters?

Steve

[Gabytrader]

The filter :
Uses the same risk as in MTP as when you enter the trade.
It deal ONLY with the bad trades!It doesn't alter the good trades! I repeat: only the bad trades !
Exactly what you heard!
It looks like the double stage filtering allow a 65% reduction of the losses!
Only Tony could help me to provide me with the rest of 170 bad trades of the last 26 weeks!
I managed only 31 bad trades!
I'll disclose all ONLY after doing all the statistic! ( the backtesting of the month-to month statistic repeatability of the last month performance!
Only then I can say it is a good method! :)
Gabytrader

[Gabytrader]

Hy, everybody!
since two weeks ago, I managed to check about 50 bad trades.
Results : the same as for the first 30 bad trades.
That is : about 65% bad trades were filtered out!
If nobody help me in sending me more Bad Trades, this Monday i'll give up with my confirming work!
Gabytrader

[stevej]

Gaby,

I've sent you a PM. Let me know if I can assist.

SJ

[SimonH]

Hi Gaby - you should have an e-mail with my few losing trades.

I think 65% elimination is a pretty good result, and wouldn't expect too much more (though if it gets better, all to the good!). The reason I say this is many programs that use a variety of methods (e.g neural networks) etc. seem to come up against this boundary, and if they get more sophisticated or over-trained they actually start losing the ability to 'see the wood for the trees' - a bit like over-zealous Elliot-wave traders not sticking to a simple ABC predictor!

Since you are interested in forex, I'll give an example in this context. All last year you could bet on USDxxx shifts based on certain significant data - like nonfarm payrolls, confidence figures, PMI, etc. Yet the last 2-3 months, despite consistent USD-positive data, the dollar persistently declined. No-one took a blind bit of notice of the 'significant' figures. So a 'system' would have been OK 75% of the time - but not 25% for no apparent reason and lots of after the event hypotheses.

I mention this also to suggest that if you do want to trade forex using MTP, then it is useful to be aware of the significant data events upcoming, and also review commentators comments on whether the market is positioning a particular expectation, as this will indicate which side of the trade carries a greater risk (or opportunity!) of large change. Note also Steve's comment that these large jumps can be short-lived, giving back most of the ground within the hour.

simon

[Steve Griffiths]

Hi Guys,

Another good Filter is to use the higher time frame, for example the 60min chart and then only trade in the direction of the 20 peroid moving avearge. This will improve results significantly.

I am just adding a new Chapter in the Trading course on this for all customers.

Thanks

Steve

[Matt Bowen]

Almost all traders (early in their trading career) shoot for 85 to 90% winners.
The statistic itself is of little value (It's probably important to their ego or
psychological comfort zone), but it really means nothing unless it is compared
to the Win/Loss ratio. The combination of these two numbers tells you all
you need to know about the logic and the room for error in the methodology.

If the system is 50% correct, the average win/loss must be greater than 1.00
to breakeven. In other words, the higher the winning percentage, the lower the
win/loss has ratio has to be to break even. Look at the matrix below:

http://users.adelphia.net/~spoonaz/Systest.gif

The next time you are looking at ANY system, you need two numbers:
1.) The Percentage of Profitable Trades
and
2.) The Average win/loss ratio
Once you have these two pieces of information you can quickly check to see
if the system in question will SINK or FLOAT...

Hope this helps,

Matt

[lezerdking]

You-If nobody help me in sending me more Bad Trades, this Monday i'll give up with my confirming work!
Gabytrader

Me- I understand it would be nice to have all the trades to confirm, and there is no reason why you shouldn't have them but, If your culling 60%+ of your big losers why wouldn't you continue your confirming work as they unfold? :(

Regards,
J.

[Matt Bowen]

Possibly left the board as well. I'm just re-listing some of these old messages so the new traders can read them.