Parabolic SAR

[wes7763]

I currently have a strategy with the Parabolic SAR. It simply enters when parabolic turns bullish and exits when it turns bearish. The problem is that I have too many whipsaws which is cutting into profits. I have noticed a trend that minimize my losses. I would like to exit if or when the ParabolicSAR turns bearish, but only if it has five consecutive bearish dots moving at an angle steeper than -45 degree angle (steeper than a slope of -1). If that criteria isn't met I want to exit when it turn back from bearish to bullish, if the bullish Parabolic Sar has five consecutive dots than are moving at less than a 15 degree angle. If neither conditions are met then I want to stay in my long position. Thanks for the help.

[NinjaTrader_Austin]

Wes, hello again. Unfortunately this is moving towards us programming for you, which is beyond the level of support we provide. Maybe some community members can chime in and provide you with some assistance.

If you'd like a professional solution, you could contact one of our 3rd party NinjaScript consultant.

[epcompany]

Wes,
You didn't ask a question but I'll send you some advice anyway.

The PSAR "looks" good as a trading tool but over the long run will run negative profit as you are seeing. It can be combined with other indicators but you tend to just break even at best. If you like PSAR, then you will fall over smiling when you look at TSSupertrend.

What I have found through extensive back testing is that is that most single or dual indicator strategies fail. If you analyse the market, you'll find that it is a random process over the long term. It also has the character of a fractal: whatever scale you look at it, it looks the same.

The trick is to model the price movement so that you get in and out with a profit. It can be done but it takes a lot of thought, time and testing. You wind up dreaming price patterns. There is a good saying: if you have a winner, don't share it because if everyone uses it, it will stop working. So, I'll just leave you with this: I use multiple indicators from NT and others on this site. I've developed other very powerful tools myself. Along the way, I needed to upgrade my programming skills from Fortran to CSharp. I get in at a "momentus" moment and get out as the moment fades. The trick is knowing when is momentus and when is not and when to get out of Dodge.

So, there is no easy way to make money here. Nobody will hand you an indicator and you trade up to your pie-in-the-sky. It will take you dedicated work over an extensive period of time. It can pay off. My method back tested over about 50 14 hour days of data (700 trading hours) makes about $500 per day single contract. When I started this using rudimentary tools like the PSAR, I lost over the long run. Simple systems can get into loosing oscillations where you can lose thousands of $$ in a single day.

BTW: Don't rely on historical data for back testing as the real market is much different. Record long runs of data and use those. Take a look at the FDAX if you like big boy games. It plays at full steam for 14 hours straight.

BTWBTW: I don't use either PSAR or TSSupertrend now.

Larry

[wes7763]

Thanks for the advice Larry. I've come to the conclusion that the parabolic SAR does seem to good to be true. I'm a rather new trader, and while I have seen success starting off, I feel like it has probably been beginners luck. I figure I need to find a system that works for me. So I have been back testing strategies trying to figure out what works and doesn't work. It sounds like you've been in my position before. The market just seems to do what it wants whenever it wants to. I am aware that real market is different than historical test, but I am curious why you say to not use historical data, but rather record long runs of data. Wouldn't that essential be the same data, and therefore could be classifed as historical data? Thanks for the advice.

Wes