Journal related to: Spud's MTF Stochastics

This is a good example of the problem with trailing Stop.

Set it too close one get stopped out too often (maybe even with slippage)

set it a little bit further - one gives away a lot of the profit.



Anyway all together good it was a good night.

Of course this is a especial problem for systems like I use with gearing just for a couple of pips per trade.

One solution to the dilema of too close and too far, reminds me of Goldilocks and the three bears,
Have two or more positions, or a position large enough to scale out.
There are arguments against scaling but I wont go into that here.
If you find yourself having too tight of a stop then it could be because you want to be sure to have some profit so act a bit chicken.
On the other hand having too big of a stop could be greed to see it run a far far ways.
So close half at say 10 pips. One can make a living at 10 pips a day you know. That way no matter what on remainding you got 10. Perhaps, depends on your method and averages. Move stop to BE -5 or BE -2 or BE+1. Somewhere that gives the last half room to breath initially and establish itself into the trend. Once up some reasonable amount for that pair, you increment the stop but not tight. Only after a decent run and it looks like it is running out of steam do you tighten with closer stops. Perhaps for less volatile the trailer is 25 pips. Once you are up 50 or more then tighten up the trailer amount.
It can depend on your strategy, time frames, etc. How far are the avg runs after your entry? Or when price approaches chart structure of some sort , then tighten up the stop trail size.

Not all trades can produce decent runs. The more swings at bat you take the more likely you wil hit a few home runs but the key is to be in at the right time to get those chances.

Why not split the trade? If you rmethod has 80%+ win rate and most wins go 40 pips then having say a 10 pip stop could be too tight if you avg 12 pips on your wins. You are missing out more often than not.

Hope that helps
Dave

Thanks mooselover.

I have actually a quite decent Trailing Stop system ( I explained it a bit somewhere in Spud's thread).

Problem here is that this are small pips 5-15 but sometimes much more.

I have also a decent system to split trades in multiple sub trades with different TP settings.

I will see how it runs this week (depending on the number of trades) - might be an option to try implementing a few sub trades with different trailing stops or one smaller TP and a other trailing from there.

Till now I never could be convinced that Trailing really produces more than a reasonable TP. At least not in my experience.


Thanks, are you still playing with MTF stochs?

The fundamental question could be, do you want to capture a set amount, presumably smallish compared to potential or do you want to capture a large chunk of a run? If you want to capture a run then some form of trailer is required. I guess one could scalp everything and never take a large bite. Alternatively enter on scalps but protect equity as soon as reasonable then see if it will run and give you a treat. You cannot predict which entries will go zoom but for sure if you always exit fast you will hardly ever catch a big run.

I am focusing heavily on TKPOWER8 method. Modfied, wrote some indi to help out and working on another now. Great goup of people and am finding success there. It does use a stoch on one of the time frames. Once am mroe settled with it all I will try adding some MTF stochs to see if it helps or hurts.
Dave

> am finding success there.

That's properly the most important thing in trading;-)

Wish you all the best.

New configured Bots work nicely (but the market moves also where excellent for this strategy)


I'm back to programming.

I got some new ideas for a way which might work out better than the trailing stop testing (which is not doing to well till now).

As I try to calculate probabilities of a trades success (which I use as additional argument for Unit Size) I thought I go a step further an adjust the TP also in a similar matter.

Not sure if that really will work but I give it a try.


From what I can tell I'm moving in the right direction with getting this system into profitable territorial.


And there is quite some possibilities of adjustments (one reason I liked Spud's extended different approaches to the same underlying Idea a lot).

Of course there is also the danger to tune it too much for a certain kind of market condition - and loose it when this changes.

funny how it goes. As soon as I had written this above I got bum.bum,bum 3 times stopped out in a row. (not on all accounts)

Still struggling.

My probability calculation is not too bad -- but it gets still sometimes very messed up as seen in the next pic.

second long trade had a probability calculated of: 63.1067961165 % LOSS

where as the First short one had a probability calculated of: 51.4563106796 % profit could be a nice run



As I use also this calculated probability to adjust a the Unit Size this is sometimes like in this case a problem.

Also not sure if it will really work out to be used for adjusting the TP setting.


Will properly make a test with no Unit Size adjustments to see what is the real different for good or bad.


Note: this calculated probability has nothing to do with the Entry --- it is just used when I enter to get some parameters to play with.

Anyway all together it works not bad (this is one of the occasionally messed up once which is still bothering me) Since the new start it looks still much better then before.


only 2 accounts in minus

Just to show the overall effectiveness of my probability calculation:

Trades Over RISK_UnitsAdjustmentPercentage: 0 --- ALLOver_Level0: 41 --> TP: 35, SL: 6: TP/Loss Ration Percentage: TP: 85.3658536585 | SL: 14.6341463415
Trades Over RISK_UnitsAdjustmentPercentage: 30 --- ALLOver_Level1: 41 --> TP: 35, SL: 6: TP/Loss Ration Percentage: TP: 85.3658536585 | SL: 14.6341463415
Trades Over RISK_UnitsAdjustmentPercentage: 40 --- ALLOver_Level2: 36 --> TP: 31, SL: 5: TP/Loss Ration Percentage: TP: 86.1111111111 | SL: 13.8888888889
Trades Over RISK_UnitsAdjustmentPercentage: 50 --- ALLOver_Level3: 26 --> TP: 24, SL: 2: TP/Loss Ration Percentage: TP: 92.3076923077 | SL: 7.69230769231
Trades Over RISK_UnitsAdjustmentPercentage: 55 --- ALLOver_Level4: 20 --> TP: 19, SL: 1: TP/Loss Ration Percentage: TP: 95.0 | SL: 5.0
Trades Over RISK_UnitsAdjustmentPercentage: 60 --- ALLOver_Level5: 15 --> TP: 15, SL: 0: TP/Loss Ration Percentage: TP: 100.0 | SL: 0.0
Trades Over RISK_UnitsAdjustmentPercentage: 65 --- ALLOver_Level6: 11 --> TP: 11, SL: 0: TP/Loss Ration Percentage: TP: 100.0 | SL: 0.0
Trades Over RISK_UnitsAdjustmentPercentage: 70 --- ALLOver_Level7: 6 --> TP: 6, SL: 0: TP/Loss Ration Percentage: TP: 100.0 | SL: 0.0
Trades Over RISK_UnitsAdjustmentPercentage: 75 --- ALLOver_Level8: 3 --> TP: 3, SL: 0: TP/Loss Ration Percentage: TP: 100.0 | SL: 0.0
Trades Over RISK_UnitsAdjustmentPercentage: 80 --- ALLOver_Level9: 1 --> TP: 1, SL: 0: TP/Loss Ration Percentage: TP: 100.0 | SL: 0.0
Trades Over RISK_UnitsAdjustmentPercentage: 85 --- ALLOver_Level10: 0 --> TP: 0, SL: 0: TP/Loss Ration Percentage: TP: 0 | SL: 0


Well this uses just 41 trades -- but the success rate improves is much higher with higher probability.

I actually did something similar last week with more trades:

Above 75 % probability was excellent but very few trades: better reasonable ration was 55 % ( actually due to the fewer trades between 55 and 75 the ration changed not always for better)



Entry signals are produced differently - but in some bots I use that as filter:

for instance (this bots have different settings from the example above)

the current 75% filter has still a 100% win ration.

But the one with not filter has produced more profit because of many more trades.

GMT -5

THIS TRADING PERIOD

START Balance: 100000.00
END Balance: 104447.24

Trades: 3
Profit: 3 | Loss: 0

TOTAL TRADING PERIOD

START Balance: 100000.00
END Balance: 104447.24

Trades: 3
Profit: 3 | Loss: 0

do small adjustments to my Bots configuration for next week.

The One I track here will also be slightly re-adjusted.

Read a nice post:


The overall purpose of this journal has changed: as I'm the only one who is left. So there is no joined effort of learning from one another.

Now the sole Purpose of this Journal is:

To show transparently over a certain time period that a defined system based on Spud's MTF presentation can be profitable.

1. To achieve this transparency I will post from time to time the complete transaction log.

2. As the purpose is only to show transparently the profitability - I will not go into exact setups as there is enough basic info on Spud's threads as well on my different postings here.

3. To make this process a bit more fun I joined with that particular account a few contests.

4. If I fail (I'm not sure yet) that would not mean that Spud's MTF ideas are wrong - it would only mean that I could not get it work or saw the continuing effort not worthwhile.

Hi M_j, Good luck in the contests. I hope you be the winner

Btw, thanks for showing me on how to convert statement to an image. It has been of great help to me

This week was all together a bad week for my test bots. Will slightly adjust for next week.



GMT -5

THIS TRADING PERIOD

START Balance: 104447.24
END Balance: 102651.11

Trades: 7
Profit: 4 | Loss: 3

TOTAL TRADING PERIOD: START 28.January 2009

START Balance: 100000.00
END Balance: 102651.11

Trades: 10
Profit: 7 | Loss: 3

Fantastic results there MJ well done

Thanks CindyXXXX.

Actually the last 2 weeks where not good at all on a general basis for the 30+ accounts.

But it is also good to have this phases now - as it exposes some of the weaknesses.

On the other hand with some of my money management applied - losses have already been reduced.

At the moment I run 45 Test accounts related to Spud's MTF.

CindyXXXX suggested to me to look into Advanced Inside Bar trading:

As most of my current projects are running without too much effort and a implantation of an automatized Advanced Inside Bar (IB) Trading seems not a too great effort (at least at the moment - who knows when I start with it
I decided to do it in a similar open style as this journal journal.:http://www.forexfactory.com/showthread.php?p=2533801

One of the nice trades

Well, this is one of the account which did very well the last view days.
(Note this is not the account which I normally display here)

Just showing how some of the others run (when they run well) -- this is much less filtered and has as such many more trades compared to my tracked account.

Restarted on Tuesday 10th with 100000.
Current: 108037.08

WIN: 28 | LOSS: 5



PS: because of the many trades I just post the TP and SL hits.



I will post later the main Account week report.

This week was all together a good week for my test bots. see post above.

THIS TRADING PERIOD

START Balance: 102651.11
END Balance: 102651.11

Trades: 8
Profit: 7 | Loss: 1

TOTAL TRADING PERIOD: START 28.January 2009

START Balance: 100000.00
END Balance: 103634.06

Trades: 18
Profit: 14 | Loss: 4