Spud's Predictive MTF Stochastics

Here are my template and indicators of Spud Dynamite (Price Chart Time Frame for 5 mins). I love this method because it is very difficult to go mechanical :

Thanks Harry123.

I have to admit I'm still not convinced about the Filter stuff.

Do you have any comparison to only using the MTF Stochastic - How much does the Filter improve the Entry -- and how much did it actually hinder a good entry or changed a good entry to a worse one ?

I have convinced myself with demo trade that this is dynamite stuff - easily improve entry accuracy by 95%. Please do the same to convince yourself. I am afraid I cannot help you in this area other than inventor, Spud, himself. OK guys/gals, I have live trades to attend to .. adios...

Last post. This has turn into a monster trade.

spud what s your opinion about the 8-17-9 macd that harry uses?

i need the fast stochastic indicator for its sensitivity. Not the slow stoc. So there is no confusion you have a fast stoch and a slow stoc. I need the fast stoch. Not different settings for the slow stoch

thanks

i've found it here it is for those who can use it

Now i see where the problem is. I can upload only .TPL files, if I upload .ex4 and .mg4 I got only this graph:

I think there may be some words missing here.

well, anyone interesting in my programming process.

It seems I managed to program a major stability improvement.

At the moment I'm in the first serious forward test phase (still might have some bugs - but at least it looks like it does what it is supposed to do)http://www.forexfactory.com/images/icons/icon14.gif



In this first forward test phase: I have 36 test bots running.

* I track 4 pairs: EUR/USD, USD/CHF, AUD/JPY, EUR/CAD

* for each pair I have 9 test bots with different settings (some are just minor others are major different settings)

* hopefully I will keep a proper statistic
(I have to do that manually for all 36 bots)
(this manual checks are important so to see the difference and not just have some numbers)

* on the Statistic sheet I will keep track for each bot: TakeProfit or other close in profit, StopLoss or other close in loss

* I still do not let my bots auto restart after an error (to better debug them) so over night for instance there might be some hanging for a while even hours which could make the statistic not accurate.

Well this is where I'm in my implementation of Spuds MTF-Stochastics.

HAPPY TRADING to all of you

MJ

Thanks for the update MJ - Been watching all of your posts to see how this amazing task pans out.

Wow

EXCELLENT work MJ! Im anxious to see more on your progress

But I have to ask about the 5 pip TP and 100 pip SL. Can't you find a more favorable risk:reward ratio?

Thanks CindyXXXX.




hi ayaz,

you are completely right: 5 pips to 100 SL is too bad.

The fact is I'm still working on the entry ;-)

Well, until recently I was a bit tossed around by new mechanical rules and other ideas. Now with my highly configurable Bot version 2 and the improved stability checks I'm starting to be able to collect statistics.



So basically I plan to proceed like this:

* let my 36 bots run day and night - they should do some work (I worked on Spud's MTF-implementation already about 1 month - approximately 5-8 hours each day)

* my main job will be: to manually check the trades on the charts - to get a feeling if this are good entries or bad once:

* check which settings are most fruitful (well here I have so many options so it might take a while)

* also it will be interesting if it works the same across this four pairs. ( properly not as my EUR/CAD Spread is much higher compaired to EUR/USD etc.)


** and after this all works fine -- I will have to look about exit rules:
I have already implemented a nice fix pips TP system for splitting each trade in as many multiple sub trades with different TP settings as I wich.

But I will also still check out if there are not better ways to exit: base on the overall Situation ov my 5 multi Time frames.

Not yet sure ?

* I also implemented already the StopLoss setting suggested by Spud himself: about 10 Pips below/above the PriceLine peak Price. - just don't use it yet.


* one feature I think might really work out positively is my UnitsSize adjustment system - base on the properpility of anicipated success of a trade.


* also I have in my overall trading program already for more than a year a couple of nice MoneyManagement systems I just have to enable in the end when I'm set.


* bottom line is still: I'm not sure if in the end the win - loos ration will be worth keeping it. with my present settings ( which I will change later) I would need 20 Winning trades for one looser which means 95 % success to break even which I doubt to succed.

Anyway - I thing my time horizont to go live with it in about 3 month is still manageable if I will progress as nicely as the last month.


If any of you is interested I can post from day to day my simple statistic I collect.

MJ


The chart below shows two trades - both where eventually Profit trades - but you see : the first was for me not a good entry - the second was one I like. Exactly for such things I'm looking now manually.

The real cool thing on this two trade is: my UnitsSize adjustment system work quite well and entered for the 1 trade the not so good entry: just 336786 Units and for the second which I consider a good one: 645921 Units.

(Well I test on a 100k Demo with purposely high percentage per trade - so if it fails one is shocked straight away ;-)

It's holiday time!

I'll be back in the New Year, wish you all the best and happy holidays!

See you in 2009!!

Spud wish you too all the best for the New Year and Happy Holiday !!!

Thanks once more MJ

20081219

Trading Session: 8:33:00 - 13:15:00
Trading Session: 13:48:00 - 19:00:00

Total 36 bots over 4 pairs (EURUSD, USDCHF, AUDJPY, EURCAD):

* TP: 146 / SL: 33

Percentage:

* TP: 81.56 % / SL: 18.44 %

Overall WORST system: based on the mechanical rules:
M5 Entries:
5 PIP TP or close by present and next higher TF-group Risk
assessment 60% or more (SL 100 PIPS)

* TP: 23 / SL: 15 || TP: 60.53 % / SL: 39.47 %

Overall BEST system: base on 16 Essential Risk assessments:
M1 Entries:
3 PIP TP (SL 60 PIPS)

* TP: 20 / SL: 1 || TP: 95.24 % / SL: 4.76 %

Overall WORST Pair: EURUSD

* TP: 45 / SL: 17 || TP: 72.58 % / SL: 27.42 %

Overall BEST Pair: AUDJPY

* TP: 30 / SL: 2 || TP: 93.75 % / SL: 6.25 %

As this are averages: from the 36 Bots 20 did make have a TP ratio of 100% (some just one trade 9 trades with 100 % TP was the highest)


As a summary: one can not say to much as this was just the first day: and from the 179 trades one has to consider that I use 9 bots pro pair and some of the trades where quite the same.

Based on this data: the overall performance would be a disaster. (this set-up needs about 96 % success)

Having said that - there is a lot of good signs with more than 50 % of the bots succeeded in 100 % TP.

SL settings can be a lot adjusted which will reduce the currently needed success rate of about 96 % to much less.


CONSEQUENCES: over night I decided to program a additional version (basically extend the last one: with new Risk assessment options - especially the Rope implementation not only on the first time frame level (as I have it now) but for the TimeFrame Level 1,2,3,4 as well)

On this new version I will drop TimeFrame level 5 all together.


* also if on Monday the results are similar - I will drop the system based on the mechanical rules (reason - with the current bot I can simulate the mechanical entry to some extend - and the TF-group switching makes for me not much sense with a small fixed TP setting. This would be much more important if I would focus on Escalators to PIP - and I can always resurrect it in the future if I need too.)



Hope some of you find it interesting to experience yourself with different settings and Pairs - remember on this particular day EUR/USD did perform worst (with my config rules) even though it has the lowest spread)

MJ

Properly all of you anyway know that - but for some beginners with MT4 like me I just wanted to give a hint how I create extended templates like Spud's Rope.

Well basically I add just one indicator of the 18 Stochastic and then I edit it manually - with past and copy and replace just change the k% or collors. I find this much faster than doing it on the graphical interface.

Well, I don't really know why I get so involved with Spud's systems.

* But somehow I loved his presentations enough to believe it 'could really work out well' (but this has still to be proven for me ;-)

* The other thing I like a lot - there is enough infos here to play with the main concept in many different ways - which gives plenty of option for Money Management and other things.

The feeling of satisfaction !!

The new additional version of my Bot 3 - is more or less ready for testing. (was much easier this time - as I could build on all the previous related stuff and had already most of the ideas in my mind)

It's huge, highly configurable and at a level it would be very difficult to maintain by manual trading. (well manually has other great levels which the bot does not have like the 'brain indicator').

ayaz will be glade to hear.

* In this version I also extended my Fix TP option to 3 levels (each with possible multiple trades)



About Money Management: The Base UNIT Size is based on the result from one of my existing extended Money Management systems.

This is completely different than Spuds 20 PIPS a day plus Pips Bank.



Base UNIT Size calculation

at the new Bot 3 I have 72 Short and 72 Long possible Risk Checks - each check can have different count weight so to give certain Checks more importance than others if wanted.

Depending on what checks are included I calculate the highest possible Risk.

my FINAL Base UNIT Size is the: Base UNIT Size multiplied by a RISK_UnitsAdjustmentPercentage
(which in turn is calculated from the Actual Current Risk Percentage of the highest possible Risk.)

Based on this FINAL Base UNIT Size I split a trade in unlimited (3, 10 , 20) configurable multiple sub trades



So a possible TP scenario could be like that:

* ESSENTIALS are met: I split a FINAL Base UNIT Size into 3 multiple sub trades

1 - TP 5 PIPS (50 % Units)
2 - TP 7 PIPS (30 % Units)
3 - TP 9 PIPS (20 % Units)

* ESSENTIALS + ESSENTIALS BONUS 1 are met: I split a FINAL Base UNIT Size into 3 multiple sub trades

1 - TP 10 PIPS (50 % Units)
2 - TP 13 PIPS (30 % Units)
3 - TP 16 PIPS (20 % Units)

* ESSENTIALS + ESSENTIALS BONUS 1 + ESSENTIALS BONUS 2 are met: I split a FINAL Base UNIT Size into 3 multiple sub trades (depending on the config this could be a Escalator to PIPs)

1 - TP 15 PIPS (50 % Units)
2 - TP 25 PIPS (30 % Units)
3 - TP 40 PIPS (20 % Units)

I will also do a Trailing SL or at least one to set Break Even for the higher TP.





So all together I feel satisfied with this concept .http://www.forexfactory.com/images/icons/icon7.gifhttp://www.forexfactory.com/images/icons/icon14.gif

Especially the MoneyManagement, TP and hundreds of configuration options.



Now - I think I will take some programming rest -- just check for stability and take care of the possible bugs (which by this huge code is quite expected)

And collect some new statistics.


If anyone is interested: I thing I shared a great deal of my system in this thread - and hope it inspires programmer as well as even none programmer to explore more of the possibilities.http://www.forexfactory.com/images/icons/icon12.gif

Basically I think - I more or less arrived in the implementation of MTF Stochastic into a fully automatized trading system.

(there might be small additions: have some ideas like including the angle of the rope etc but not in the immediate future)




Now it's time to forget new stuff (for instance I do not allow for elasticity price lines (opposite of the Origin or Rope peak)) - fix remaining bugs and experiment with different configurations.



And most of all: Prove if it is overall well profitable - and get live with it in about 3 month time -- otherwise just get rid of it. (well at least it should pay me 2 month wages I will have invested in the end)http://www.forexfactory.com/images/icons/icon12.gif

So that should be all on the technical side of my MTF Stochastic project - as most members are not programmers - I do not want to spoil Spud's excellent thread.

Once more THANKS to all who contributed and especially to Spudfyre to share this vast information in a very helpful way.

MJ

MJ,

What is bot?

http://en.wikipedia.org/wiki/Internet_bot

hope that helps.