Anyone Making money???

[Trevor.Richter]

These are price action targets and entry signals that were determined realtime. Just a couple of recent examples. This method works. More later.

Which method would that be?

[eDanny]

Yesterdays chart. More later...

[modernmind]

comment on 1/3 of 10 cents

Most interesting thread dudes!

Trading is difficult because 99% of us are wired up the same - hence we all fight with each other to get the same trades.

Whether its RSI or Stochastics there isn't a huge amount of difference.

For me the holy grail consists of 3 things

1. write your plan down and have absolute discipline - still working on that
2. keep constant and detailed records
3. work out your position sizing - I make more money playing with that than anything else

I employ 3 indicators - a moving average oscillator and a lin reg oscillator and an elliot wave type indicator. Half the time I forget that I have any of them!

To me - trading is like a box of chocolates you pull the lever and never know what you are going to get on THIS trade. If you have an edge overall then you MUST have faith over a number of trades.

Hello Mindset,
I wonder if you could elaborate on your personal position sizing method. I've read a lot of posts on this subject and it seems that the "built in" method for position sizing in NT6.5 is not ready for prime time due to the fact that it doesn't adjust dynamically to the current account balance. I see that there is an add-on available via Positionsizer.com that seems worthwhile (but very expensive). I haven't heard anything mentioned with regard to better position sizing support for NT7. Have you developed an efficient method for position sizing other than a spreadsheet?

[Mindset]

modernmind

There are 2 elements to my size:

1. I trade 1% of equity which is updated daily per trade.
2. I trade multiples of this depending upon the quality of the signal. My definition of quality is the signals actual expectancy over a large no of trades.

In terms of coding I just use

NB pseudo code to aid explanation - just alter to suit
double riskEquity = "x no of dollars";// or Account equity * 1%,etc
double riskDollar = Math.Abs(entry price - stopprice) * Bars.Instrument.MasterInstrument.PointValue;

positionSize = ((riskEquity / riskDollar);

(if you want the size in whole nos substitute below for above  

positionSize = (((int)( riskEquity / riskDollar));

[Mindset]

position sizer

I don't use this - I have my own version - but that really is good value for money. Sizing is the holy grail IMHO.
There is some great coding in there and if someone ever shows me the way to get things onto the toolbar - then away we go!

Edit - having trialled it for a bit there are a few little bugs in it. The ability to trade the position though direct from the toolbar is very cool.
for me there is still way too much clutter on the chart - but that's personal.
Maybe there will be some toolbar button help in NT7?

[heech]

I was VERY tempted to post here at the end of August, crowing about my success... and it's a good thing I didn't, because I gave it all back in September.

I was +2.8% in July, and +4.2% in August. September, I gave back the August winnings... and I'm down (marked to market) about 5% on positions expiring in October.

All in all, a *very* mixed bag. It's hard to say what will be more common, August or September... obviously I'm hoping for the former!

I will say one thing though: NinjaTrader is working very well for me. I'm basically starting the program once a week Sunday afternoon, launching my strategy, and then that's it.. basically hands off for the next 5 days, fully automated. I'm doing probably 200 orders a month through Ninja, and they're working well. There are still a few exceedingly rare race conditions that are just about impossible to duplicate, so I still have to be on my toes occasionally... but things are going far better than I expected at this stage.

[cunparis]

I was VERY tempted to post here at the end of August, crowing about my success... and it's a good thing I didn't, because I gave it all back in September.

I was +2.8% in July, and +4.2% in August. September, I gave back the August winnings... and I'm down (marked to market) about 5% on positions expiring in October.

Maybe you curve-fitted your system?

I haven't had an automated system hold up over the long run.

[heech]

Maybe you curve-fitted your system?

I haven't had an automated system hold up over the long run.

I'll let you know in about 12 months? Haha.

[cunparis]

I'll let you know in about 12 months? Haha.

Do you use indicators? Did you optimize your strategy to find the "optimal" parameters for the indicators?

To update this thread for me, 4 weeks ago I tried something different. I removed all indicators and started trading price & volume only. Took me a few weeks to get used to it but I'm now starting to be profitable. Never going back to indicators again. But I'm doing this manually, not automated..

[MicroAl]

heech

Glad to know you are making progress. I have decided to go back to system trading for one of the advantages you mentioned, the ability to let the system run on its own. I would rather spend time programming and investigating setups rather than spend all day watching a screen.

Also now that with Ninja 7 Multi-instrument graphs etc are now finally possible I should finally be able to do the things with Ninja that I wanted to do for the last 2 years!

[heech]

Do you use indicators? Did you optimize your strategy to find the "optimal" parameters for the indicators?

To update this thread for me, 4 weeks ago I tried something different. I removed all indicators and started trading price & volume only. Took me a few weeks to get used to it but I'm now starting to be profitable. Never going back to indicators again. But I'm doing this manually, not automated..

Just to update again... the October positions that had been negative (marked to market) are now very positive, and I'm looking at +7% for the month. Very pleased about it.

I'm now back into profitable territory... and on the basis of actual values at expiration, I haven't had a negative month since June. I'm up an average of 2% a month since then.

[nicknamed]

system

Maybe you curve-fitted your system?

I haven't had an automated system hold up over the long run.

to avoid curve fitting a good system should perform well on all non-random walk data which currencies and futures are, so if you have stable system that makes money on currencies, it should be able to make money also on any other asset class - vice versa (especially on currencies trends are strong, so good system must make there a fortune), though every asset is behaving differently so somewhere it should make more money and somewhere less or it may take longer to make that profit but overall it should work on any timeframe and asset class (if u look on profit graphs after a year, they all should have the same pattern). if it does then it should perform well over long term.

to do that I recommend: use standard values for your indicators, optimize system on artificial data series where you mix many asset classes - u might need to have continuous data series, and then check profit graph and if it looks about the same over different assets, go realtime....

[sefstrat]

There is some truth to what you say in that a stable system should be at least profitable on any asset class. However the most profitable system (assuming one which trades only a single instrument, not pairs trading etc) will always be one tailored to the specifics of that instrument, or in other words 'curve fitted'.

For instance there are observable and mathematically quantifiable differences between say EURUSD and USDJPY, relating to factors such as time zone differences and the heavier volume of trading on EURUSD among other things.

If your strategy is optimized to take advantage of such things it is technically 'curve fitted'. I guess what I am trying to say is that optimization/curve fitting is not inherently bad, quite the opposite in fact. However, designing a strategy and using optimization properly is difficult, therefore many traders say 'curve fitting' as if it were a dirty word only because they do not understand how to do it properly.

Optimization means you are taking something good and tweaking it to make it even better. It is not a solution that excuses you from having to think about the problem. In other words, your strategy should be profitable across a wide range of instruments with no optimization whatsoever, none. Only then is it safe to add optimization into the mix. If your strategy is not profitable until you run it through the optimizer to find the right parameters then it is poorly designed and the likelihood of it being stable in the long term is extremely slim.

to avoid curve fitting a good system should perform well on all non-random walk data which currencies and futures are, so if you have stable system that makes money on currencies, it should be able to make money also on any other asset class - vice versa (especially on currencies trends are strong, so good system must make there a fortune), though every asset is behaving differently so somewhere it should make more money and somewhere less or it may take longer to make that profit but overall it should work on any timeframe and asset class (if u look on profit graphs after a year, they all should have the same pattern). if it does then it should perform well over long term.

to do that I recommend: use standard values for your indicators, optimize system on artificial data series where you mix many asset classes - u might need to have continuous data series, and then check profit graph and if it looks about the same over different assets, go realtime....

[nicknamed]

"If your strategy is optimized to take advantage of such things it is technically 'curve fitted'. I guess what I am trying to say is that optimization/curve fitting is not inherently bad, quite the opposite in fact."

I cannot agree with this paragraph. I spent a year with neural networks and I perfectly know that curve fitting is bad in general, no excuses and one has to avoid it as much as possible - in other words the more you avoid it, the longer your autmatical system will make money, though less thn when fitted well but working for shorter time. BUT you are right indeed just in the next paragraph saying

... your strategy should be profitable across a wide range of instruments with no optimization whatsoever, none. Only then is it safe to add optimization into the mix.....

that is true, I can sign under that, say my system right now was constructed with standard setting and u know what when I tried to optimize, the results vere worse on many instruments...very interesting indeed, one would expect something different....