A stochastic Oscillator system

[jezza888]

Hi all,

I've been reading through this site and not found quite the answer I was looking for and would appreciate anyones thoughts on the following. (except for being told not to use indicators.)

I've tweaked the S.O. in Metatrader4 and found a succesful way of trading cable of a 15 minute chart as shown in the attatchment.

The tests go back 12 months and I've been trading it with a capital spreads demo account for two weeks now and the results are great with high success, minimal losses and a few breakevens and I'm getting ready to go live.

My only concern is, many threads on this board say about how the markets change and what worked in the past won't work in the future etc... Now, is that for all POV (P/V or indicators) or what? On one hand theres loads about backtesting and ensuring the viability of your system (ie last Knowledge lab article) yet just as much about how its pointless.

I feel confident to go live and this is my only nagging fear that all my hard work and hours testing a system is going to leave me flat on my face and out of pocket in two weeks time when things change!

The system is using the stochastic Oscilator and surely that will allways correspend with the price in the same way as its a mathematical equation regardless of the the MM and traders style and attitude?

 

[commanderco]

try running stoc(2,1) and stoch(3,1) and watch the values of each and
their interelationship.

 

[jezza888]

Cheers guys.

Joules, I am a tight git at the best of times so my money management is very good and I risk as little as possible at any given time! I use tight stops and have allowances in the system to be able to make the most of the trends and runs.

the above example isn't a one off good day, just today's, and as long as the crossover doesn't erase itself by changing direction immediately (avoided by not rushing in) with tight stops and break even stops ASAP, the crossovers themselves are an exact indication of entry and exit removing a large preportion of interpretation.

commanderco, not quite sure how you mean?

My confidence is there over the entries and exits etc but my concern is the nature of the S.O. over the future as traders change, how different and messed up can the S.O. become compared to the price movement. Over the last twelve months its been great but who can say if it won't go horribly pear shaped as soon a s I go live or two weeks down the line.

Can a mathematical euation go that far off course in relation to the price from the results produced over the last 12 months???

 

[commanderco]

jezza888

If you have been back over the 15 min charts for twelve months
and if you are happy with the output when the market is long short or sideways
then what is the problem.
12 months of 15 min action is a lot of occurences.
I trade short in a slightly different manner than long since the upside/downside volatilities are different, but that is just fine tuning.

Good luck

 

[jezza888]

Thanks Commanderco,

I know exactly what your saying... the problem is, as I have read numerous times on the boards that market behaviour changes and systems don't work after periods. My concern is that I'll go live next week and it will change so that i lose all slowly and have to start again.

I want to have that extra confidence in my belief that it will work in the future like it has in the past. To believe that as the S.O. is worked out the same way it will always show the same opportunities. Just to have that extra faith when I start putting real money on the line.

 

[ducati998]

jezza

I want to have that extra confidence in my belief that it will work in the future like it has in the past. To believe that as the S.O. is worked out the same way it will always show the same opportunities. Just to have that extra faith when I start putting real money on the line.

Lets just look at some of the language that you use,
extra confidence, belief, future, past, always, faith, real money, on the line

What this very quickly indicates is that you are not yet ready to trade real money.
Why?

1......Joules gave you the answer..............money management. If it does not work, you will find out quickly, and be able to limit the damage.

2.....If you have tested it correctly, you will know your potential drawdowns, you will know your "probabilities" or expectancy, you will know your profability ratios, and you will know your timeframe to enter profit, and you will know your equity curve, based on all the above

If you do not know these, you are guessing & hoping, and you will lose money

3.....Paper trading is a very different proposition to trading real cash, particularly cash you can ill afford to lose, or require for living expenses.

You do however need to dip your toes in the water.
Cut your bank by 75%, and go in very small to begin with.
If you do well, then you can increase, if badly, back to the drawing board.

cheers d998

 

[jezza888]

I appreciate you inpput ducati998 but I still haven't attained quite the answer I wanted to hear.

I am aware of all you said and I am ready to go live. My MM is strict and theres no hopes guesses in my system.

What I was hoping to hear was that regardless of it being last year, ten years ago, or next year, the stochastic oscilator will still work in accordance with the price in the same way. Which, the more I think about it the more I believe it should so I shall trade real money and trust what I've already proven to myself.

 

[breadman]

I appreciate you inpput ducati998 but I still haven't attained quite the answer I wanted to hear.

I am aware of all you said and I am ready to go live. My MM is strict and theres no hopes guesses in my system.

What I was hoping to hear was that regardless of it being last year, ten years ago, or next year, the stochastic oscilator will still work in accordance with the price in the same way. Which, the more I think about it the more I believe it should so I shall trade real money and trust what I've already proven to myself.

Jezza888

I have also been looking at a stochastic system in GBP/USD 60 mins over the last year the results realtime but not traded are June +107pips, July +606, Aug +1483, Sept +888, Oct -164, Nov +1182, Dec +1234 ,Jan -360, Feb +64, Mar +288, Apr +23, May -278, June +383. This year Jan to June has been a lot harded to make money. Has you system given you more losing trades in the last six months.

Thanks,
Breadman

 

[jezza888]

the first few months of the year weren't the greatest but it seemed quiet and flat compared to the end of last and the last month or so has picked up again.

In a relatively laclustre period it seems quite inneficent but still worthwhile but the overall picture is good.

whilst I can accept low amounts/movements, I don't want to find my settings change in two months time and although I could test every day/week/month back 10 / 20 years, I guess you can never be 100% sure of the future (such is life) and never guarantee the income and thats what gets me.

 

[jezza888]

It has nothing to do with if I'm ready... maybe I should paper trade until I am malarky.... another 5 days or 5 weeks of paper trading is pointless as the next day is when it could all change! the day real money is on the line!

 

[jezza888]

so much talk of backtesting and devising systems but there is no way of being suree what worked today or last week will work tomorrow, whether you use oscillators or price and volume or Gann, fibanacci or the stars. even FA can only be assured to a point.

how do you go from 'gambling' to 'a calculated systmatic trader with risk control' etc. not one direction for anticipation is guaranteed so why bother planning it, why not toss the coin and use MM and tight stops...?

 

[jezza888]

anyone have thoughts on the validity and % success rate of a system based on a oscilator/indicator then?

Yesterday is all well and good but is that a guarantee for tomorrow?

 

[blackcab]

If you're developing a system that has variable parameters on historical data, then split the data into test data and verify data. It's the best you can do with backtesting - have you been doing that?

If so, how have you been selecting the test and verify sets? Only once or many times? Randomly, systematically or by gut feel? By hand or programatically? In blocks or interleaved?

 

[FetteredChinos]

anyone have thoughts on the validity and % success rate of a system based on a oscilator/indicator then?

Yesterday is all well and good but is that a guarantee for tomorrow?

depends on your exit...

use the oscillator only for entry, and trail the stop by a multiple of the current average true range.

do not use the oscillator as an exit. only exit when stopped out by the trailing stop.

the worst you will do over time is breakeven. more likely, you will get a profit, though not particularly substantial.

it is then up to the quality of your entry to make the system more efficient...

 

[jezza888]

Would have missed alot yeaterday if TS rather than Oscillator....

Have a look yourselves...

(Metatrader4)

K% 8
D% 4
Slow 7
MA linear weighted
price field close/close
lines at 15 / 85

8am - 9pm BST
crossover to crossover
stay in until exit from outer lines (15/85)

cheers

Jezza