Anyone Making money???

[ninjablanks]

Great thread! Now with NT I am finally able to write and test my ideas. I have also tested hundreds of indicators and systems. I trade an automated system on the Euro futures. I only trades about 1 x everyday with a 45% win ratio. Target is 2 x stop loss so it is profitable (so far!) One of the hardest things I have encountered is developing the discipline to let the system run. The things that I constantly debate are:
1) Do I want a system with a high winning % (with smaller targets and stops) or do I want lower % with higher targets and stops. So far I have leaned toward Lower winning %.

2) What should target price/Stop loss ratio be? In my view exits are absolutely critical bc they determine winning percentage, how often system trades and overall profitabilty. I use fixed targets but have experminented with various types of momentum exits (RSI, MA) only to return to fixed targets bc flexible exits gave back too much open profit. The occasional big winner was not enough to offset the erosion of regular profits.

3) Anyone have any ideas on a better exit strategy?

I trade on a 15 min TF 1 contract so far I am up (90 days)

Thanks

[Operandi]

I've been been speaking with my colleague in the high frequency trading team and they are still struggling to get their HFT system to work in the currency markets at the moment.

This is a high powered and talented team with a huge amount of resources.

Also I heard about another HFT team in a major bank that was shut down last week because their systems were not working either.

So some of the big players are struggling too!

[aqs12]

TAJ,

I think I'm on the same line as your strategy.. I use use a 4 min chart instead.. it is the anchor of my trading system..

Mike

2 point Stops (Non Negotiable!)
Initial Target is 3 points. But like in a war ( and trading is a war) the plan usually is thrown out the window when the battle begins. I will not wait for a 3 point target if I see something in the price action that I do not like. And then there are times when price shots through the initial target and I just go with it. On a few occasions I can get a 10-15 point winner . I trade old school so I do not auto trade nor do I use the ATM feature. I do have alerts programmed when a potential setup exists but I may or may not take the trade. Experience has taught me when to tread lightly and every trade is a new experience. There are occasions when I will take a trade when the base setup is not there. That is the subjective part of my trading.

The 5 Min Chart is the King for me and then I use the 500 Vol Chart for timing the entry and spot the first signs of trouble. I also have 1440, 15 and 1 Min charts open. I only use three indicators, Bollinger Bands, StochasticsFast and then my own developed indicator for timing on the 500 Vol chart. I also have some Pivots and Fibs but I don't really rely on them for decision making unless they become PROVEN support or resistance. That is it; very simple. I need simple do to synapse power available LOL.

How about your TP and SL's? Charts, Indicators?

[andrewbee]

I make money, but it's not using a pure NT strategy. All my NT strategy does is write the price out to a file where a Java program picks it up, makes trading decisions based on it, and then uses the File interface to place and manage orders.

This is no reflection on NT - given enough time I could re-code it all in C#, but my investment in Java code is so great it is easier just to use NT as the front end.

[cunparis]

I make money, but it's not using a pure NT strategy. All my NT strategy does is write the price out to a file where a Java program picks it up, makes trading decisions based on it, and then uses the File interface to place and manage orders.

This is no reflection on NT - given enough time I could re-code it all in C#, but my investment in Java code is so great it is easier just to use NT as the front end.

So your java app writes out the orders to a file that NT scans and then NT places the order?

I have been using Java for a long time and I'm much more comfortable with it. I've been debating writing algorithms in Java. Up until now I've just been learning C#. The language doesn't bother me, what bothers me is not having Eclipse, debugger, JUnit, mock objects, etc. Mainly the IDE.

[andrewbee]

Correct. My Java app uses the File interface, which is a built-in NT feature. I tried to copy and paste a URL to the help section on it, but couldn't copy the URL for some reason.

The File interface is for order management. The only C# code provides a price feed to the Java app.

[Trevor.Richter]

Sage advice

I think that for me, it's not so much about the losses during discretionary trading, but just "sitting there". It's the "bored" trades that kill me.

Babypips has a great bit of advice that can be applied to hundreds of trading situations, which is "cut off your hands"...as in: when tempted to make discretionary trades even though you have an automated system that's running fine without you, cut off your hands!

"http://www.babypips.com/school/protect_yo_self_before_you_wreck_yo_self.html

[ctrlbrk]

One thing I find interesting in this thread is there are clearly many different strategies/techniques/approaches being applied here, with the same goal in mind. To be expected, yes -- but I was surprised by the shear differences in the approaches.

Another thing that comes to mind... it seems most people who've posted in this thread so far are using some pretty advanced techniques. I was not expecting this, as I figured a lot of successful traders were using very simple methodology to trade, but it seems we've each created our own rather complex system.

Mike

[heech]

This month is looking to be much more positive than last month... I was down probably about -10% on the previous option cycle, and am expecting to be up about +8% this cycle. That still leaves me net down, but I've made some significant improvements... which makes me want to offer this little insightful gem that I hope help others:

- spend as much time trying to improve your strategy, as you do on *screening* candidates for your strategy.

There are thousands of symbols out there. Some will do better with your strategy, and some will do worse. Both pieces are critical. I thought I had optimized the heck out of my strategy, only to see it fall apart completely during the rapid run-up last month. This month, I figured out that I need to be more selective in choosing the stocks I work with.

I'm hoping for all wins the rest of the year. Stay tuned.

[RJay]

One thing I find interesting in this thread is there are clearly many different strategies/techniques/approaches being applied here, with the same goal in mind. To be expected, yes -- but I was surprised by the shear differences in the approaches.

Another thing that comes to mind... it seems most people who've posted in this thread so far are using some pretty advanced techniques. I was not expecting this, as I figured a lot of successful traders were using very simple methodology to trade, but it seems we've each created our own rather complex system.

Mike

Mike,

I have been studying and learning here at NT for one year this week.

I have researched, tested everything I could get my hands on that was free. I also seriously considered buying a system.

Just to let you know, I'm a simple guy. I like things simple.

I am currently testing a couple of my own strategies with my own unique charts and indicators.

At this point they are very simple. They are 100 percent mechanical systems.

This week, trading in sim mode with my primary system.

I got my clock cleaned Monday trading ES.

Tuesday, I made a killing. Today I did "OK". Those sudden jumps in the market today were brutal.

I hope to have a consistently profitable system tested in the next couple of months.

I agree that it is best, at least for me, to design your own system.

Best Wishes,

RJay

[dave1992]

Is it true that the professionals have to trade each day due to the money each 'seat' is costing (I mean the investment banks etc not the floor traders). Perhaps someone could come back on this.

Not all of them have to trade every day. However, the rest of the time they are planning, researching etc.

Developing a winning strategy on backtesting is easy. Would take me 10 mins. None of those strategies would win on live. People who don't get this need to understand curve fitting - you fit the optimised parameters to the historic price data to maximise profit. The future price data is different so your parameters won't work.

One thing you can do is develop a strategy that would work on many different instruments. You can take account of different tick size, volume, volatility etc in the code. If that is profitable for all instruments, that's a good place to start. Your parameters need to be general things, rather than 'if my profit is now 0.81% then close'.

It's also very useful if your automated strategy can take account of extreme reactions to news. e.g. if the the price suddenly moves the wrong way over a certain amount, close!

[sefstrat]

My system fails miserably in backtesting but is quite successful in real trading. Why? Calculate on bar close == false

Test every money management strategy you can find. Learn the theories behind them, develop your own. Ever wonder why you can download a gazillion indicators but almost no money management strategies?

Use multiple timeframes.

Develop a thorough understanding of the mathematics behind your indicators and the statistical implications of using them with different bar types.

Indicators that tell you when *not* to trade are far more valuable than those which give you buy/sell signals.

I just ran a backtest on my system with EURUSD over the past 4 days (5/3-5/7). It lost over $42k with a performance ratio of .32, 1800 trades. Oh and that is not including the almost $30k in comission.

In reality it only took about 250 trades in that same period, with very high win % although about 80% of the winners were stopped out at breakeven or a few ticks beyond.

Why the disparity? Because it makes extensive use of information which is only available in real time, market depth, volume, order flow to ensure that only low risk entries are taken.

High frequency trading is possible.. but it requires a lot of work. It requires that you intimately understand the dynamics of the market and the methodologies employed by the other players. In that sense, it is a lot like reverse engineering (ever notice how many of the best automated traders are from russia?) and this is why many PhD level mathematicians will fail, because they often approach the topic from the completely wrong angle.

(Some of the above info is more applicable to FX than to index futures like ES, where there is a less dierct link between order flow and price action. ES is also much less noisy and much easier to trade however)

[RJay]

I also have horrible results with NT's backtesting.

NT consistantly puts the orders in the wrong places when backtesting.

I have to test my stuff in real time only. This is very time consuming.

RJay

[NinjaTrader_JoshP]

NinjaTrader does not place orders in the wrong places. NinjaTrader takes the most conservative approach when it comes to backtesting to prevent overinflation of results. Orders will be placed in the next feasible location. This means the open of the next bar. You cannot trade on a bar that is already closed.

Furthermore, backtesting should NEVER be the be-all, end-all testing. Backtesting provides you a small glimpse of one of the possibilities your strategy can take. You should backtest against multiple things, optimize, walk-forward optimize, backtest with different parameters, backtest with different dates, etc. all in an effort to get a fuller picture of the potential of your strategy. After all of the backtesting you absolutely need to go into forward testing. These two things go hand in hand. You should never think you can just backtest and everything will be fine to start trading live. After you've forward tested the heck out of your strategy too, only then should you even consider starting your strategy live. Even then though, you need to take ALL results with a grain of salt. Testing is absolutely 100% NEVER the same as real-time live trading, but that is why you needed to run so many tests to begin with. You want to see the wide range of fluctuations possible with your strategy and decide if that range is acceptable. Only then should you move forward with your strategy. Time consuming? Absolutely. This is why we provide NinjaTrader for free for as long as you need to finish the complete testing phase.

[RJay]

For backtesting a strategy,

When I program an order to be placed at bar open plus 5 ticks, NT placed the order at bar open and ignores the 5 tick offset.

5 ticks off of every trade during a backtest really screws up the numbers.