Short term S&R system

[nonlinear5]

Quote from RoughTrader:

Now to define the setup:

- If the high of the current bar is greater than the R_Level of the previous bar, then Setup = 1

- If the low of the current bar is less than the S_Level of the previous bar, then Setup = -1;

I tested it on the ES data from January 8, 2008 to December 23, 2008. As I expected, it turned out to be a consistent loser. Here is what I got (trading 1 contract with your default period of 20 minutes):

Trades: 3166
Percent Profitable: 34%
Net Profit: -$32160
Max DD: $44270
Profit Factor: 0.89

However, I should mention that I didn't use the profit target (you didn't describe it sufficiently), and I didn't use your other entry conditions which you chose to keep private. So, I suppose that all that my backtest proves is that buying the breakout of the 20 min high and selling the breakout of the 20 min low is not a profitable strategy. If you provide more information about what goes into your system, I'll be happy to backtest it for you.

 

[nonlinear5]

Quote from nonlinear5:

I tested it on the ES data from January 8, 2008 to December 23, 2008. As I expected, it turned out to be a consistent loser. Here is what I got (trading 1 contract with your default period of 20 minutes):

Trades: 3166
Percent Profitable: 34%
Net Profit: -$32160
Max DD: $44270
Profit Factor: 0.89

However, I should mention that I didn't use the profit target (you didn't describe it sufficiently), and I didn't use your other entry conditions which you chose to keep private. So, I suppose that all that my backtest proves is that buying the breakout of the 20 min high and selling the breakout of the 20 min low is not a profitable strategy. If you provide more information about what goes into your system, I'll be happy to backtest it for you.

For illustration purposes, here is a typical day on the chart below. Red circles indicate short entries, and green circles indicate long entries. The number inside the circle is the number of contracts (long or short). As can be seen, the first trade is profitable, but the numerous subsequent losses make it a losing day.

 

[RoughTrader]

Quote from nonlinear5:

For illustration purposes, here is a typical day on the chart below. Red circles indicate short entries, and green circles indicate long entries. The number inside the circle is the number of contracts (long or short). As can be seen, the first trade is profitable, but the numerous subsequent losses make it a losing day.

Yes, that is what I expected. Like I mentioned, using the setup as an entry will produce a losing strategy.

I can't go into any further detail other than to say that there are ways to produce the curves I have shown in this thread, using the same setup. However, If you post your code, I can see if there are differences other than the entry I use.

rt

 

[nonlinear5]

Quote from RoughTrader:

I can't go into any further detail other than to say that there are ways to produce the curves I have shown in this thread, using the same setup. However, If you post your code, I can see if there are differences other than the entry I use.

Well, if you can't discuss details, then I can't do much more for you, and I guess there is nothing left to discuss.

I use my own trading/backtesting/optimization platform, so I am not sure if my code will make sense for you, but it's pretty straightforward.

 

[RoughTrader]

Quote from nonlinear5:

Well, if you can't discuss details, then I can't do much more for you, and I guess there is nothing left to discuss.

I use my own trading platform, so I am not sure how much sense my code would do for you, but it's pretty straightforward.

nonlinear,

Please don't misunderstand me, I'm not looking for help from anyone. The point I have been trying to carry across to anyone interested in this thread is clear.

Simple setups work, but only if all components of the strategy are there to complement the setup. In this case, the entry is critical here. It literally pays to think about designing the entry apart from the setup itself. A lot of traders simply restrict themselves to using the setup as the entry, and that is a very restrictive approach that makes designing simple systems that actually work quite difficult. In this case, the simple dynamic S/R designed by the OP does just fine as a setup for a breakout system, even though it does not discriminate price geometry.

All the same, good luck to you.

rt

 

[IndexSwing]

Quote from nonlinear5:

For illustration purposes, here is a typical day on the chart below. Red circles indicate short entries, and green circles indicate long entries. The number inside the circle is the number of contracts (long or short). As can be seen, the first trade is profitable, but the numerous subsequent losses make it a losing day.

If you notice in this example we have a general downtrend with a lot of short term counter trend rallies. I think the issue in this example is that the metric used to exit is too short term. I would be interested to find out how the following rules would test out.

Enter long when the market makes a new 7 hour high. Exit long position when the market makes a new 2 hour low.

Enter short when the market makes a new 7 hour low. Exit short when the market makes a new 2 hour high.

Another variation would be to add an ATR stop.

 

[bmwhendrix]

Simpler yet, just takes longswhen price above an 80ema, and below for shorts.

 

[IndexSwing]

Quote from bmwhendrix:

Simpler yet, just takes longswhen price above an 80ema, and below for shorts.

But that does not account for S&R. I think that a markets ability to move up above a previous high or down below a previous low may be what is most important.

 

[bmwhendrix]

I agree, What I meant, was take S/R based enrties that agree with a moving average. This cuts out some whipsaws, but also prevents you from catching early reversals.

 

[IndexSwing]

Quote from bmwhendrix:

I agree, What I meant, was take S/R based enrties that agree with a moving average. This cuts out some whipsaws, but also prevents you from catching early reversals.

Oh ok gotcha.